WO2022249331A1 - 画像処理装置、画像処理方法、およびプログラム - Google Patents
画像処理装置、画像処理方法、およびプログラム Download PDFInfo
- Publication number
- WO2022249331A1 WO2022249331A1 PCT/JP2021/019993 JP2021019993W WO2022249331A1 WO 2022249331 A1 WO2022249331 A1 WO 2022249331A1 JP 2021019993 W JP2021019993 W JP 2021019993W WO 2022249331 A1 WO2022249331 A1 WO 2022249331A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- query
- image
- frame
- feature
- search
- Prior art date
Links
- 238000012545 processing Methods 0.000 title claims abstract description 212
- 238000003672 processing method Methods 0.000 title claims description 10
- 238000004364 calculation method Methods 0.000 claims abstract description 108
- 230000010354 integration Effects 0.000 claims abstract description 35
- 230000008859 change Effects 0.000 claims description 126
- 238000000034 method Methods 0.000 claims description 109
- 230000008569 process Effects 0.000 claims description 42
- 210000000988 bone and bone Anatomy 0.000 description 129
- 230000036544 posture Effects 0.000 description 106
- 238000010586 diagram Methods 0.000 description 47
- 238000001514 detection method Methods 0.000 description 43
- 230000009471 action Effects 0.000 description 31
- 210000002414 leg Anatomy 0.000 description 21
- 238000010606 normalization Methods 0.000 description 16
- 238000012986 modification Methods 0.000 description 14
- 238000005516 engineering process Methods 0.000 description 10
- 238000010801 machine learning Methods 0.000 description 10
- 230000006399 behavior Effects 0.000 description 9
- 210000003127 knee Anatomy 0.000 description 6
- 238000012544 monitoring process Methods 0.000 description 6
- 230000037237 body shape Effects 0.000 description 5
- 239000000284 extract Substances 0.000 description 5
- 230000006870 function Effects 0.000 description 5
- 238000012937 correction Methods 0.000 description 4
- 238000003384 imaging method Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 241001465754 Metazoa Species 0.000 description 2
- 230000001186 cumulative effect Effects 0.000 description 2
- 230000007423 decrease Effects 0.000 description 2
- 230000003247 decreasing effect Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000005401 electroluminescence Methods 0.000 description 2
- 210000001930 leg bone Anatomy 0.000 description 2
- 229930014626 natural product Natural products 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- FFBHFFJDDLITSX-UHFFFAOYSA-N benzyl N-[2-hydroxy-4-(3-oxomorpholin-4-yl)phenyl]carbamate Chemical compound OC1=C(NC(=O)OCC2=CC=CC=C2)C=CC(=C1)N1CCOCC1=O FFBHFFJDDLITSX-UHFFFAOYSA-N 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 235000000332 black box Nutrition 0.000 description 1
- 238000007621 cluster analysis Methods 0.000 description 1
- 239000000470 constituent Substances 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 210000003414 extremity Anatomy 0.000 description 1
- 230000001815 facial effect Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 210000002411 hand bone Anatomy 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/103—Static body considered as a whole, e.g. static pedestrian or occupant recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/48—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
- G06F16/483—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using metadata automatically derived from the content
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
- G06F16/43—Querying
- G06F16/432—Query formulation
- G06F16/434—Query formulation using image data, e.g. images, photos, pictures taken by a user
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/50—Information retrieval; Database structures therefor; File system structures therefor of still image data
- G06F16/53—Querying
- G06F16/532—Query formulation, e.g. graphical querying
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
- G06F16/73—Querying
- G06F16/732—Query formulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/74—Image or video pattern matching; Proximity measures in feature spaces
- G06V10/761—Proximity, similarity or dissimilarity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
Definitions
- the present invention relates to an image processing device, an image processing method, and a program.
- Patent Documents 1 and 2 are known.
- Japanese Patent Application Laid-Open No. 2004-200000 discloses a technique for searching for similar human poses based on key joints such as a person's head and limbs included in a depth image.
- Japanese Patent Application Laid-Open No. 2002-200000 discloses a technique of searching for similar images using posture information such as tilt added to an image, although it is not related to the posture of a person.
- Non-Patent Document 1 is known as a technique related to human skeleton estimation.
- Patent Literature 3 describes that when a reference video serving as a query is input, similar videos are searched using the number of faces of characters and the position, size, and orientation of each character's face. It is
- Patent Document 4 also describes calculating an integrated scene feature amount by averaging the scene feature amounts of each of a plurality of query videos. Furthermore, Japanese Patent Application Laid-Open No. 2002-200000 describes a technique for extracting an image that may be popular from a moving image composed of a plurality of frame images.
- One of the objects of the present invention is to enable accurate retrieval of an image showing the same posture or action as the posture or action indicated by a query.
- query acquisition means for acquiring a first query moving image having a plurality of time-series first frame images and at least one second query moving image having a plurality of time-series second frame images; feature quantity calculation means for calculating a feature quantity of a key point detected from an object included in each of the first frame image and the second frame image; correspondence identifying means for identifying the second frame images corresponding to each of the plurality of first frame images; A process of integrating the feature amounts calculated from the first frame images and the second frame images corresponding to each other is performed on a plurality of combinations of the first frame images and the second frame images corresponding to each other. By doing so, an integration means for creating an integrated query in which the integrated feature values are arranged in chronological order; is provided.
- the computer a query acquisition step of acquiring a first query moving image having a plurality of time-sequential first frame images and at least one second query moving image having a plurality of time-sequential second frame images; a feature quantity calculation step of calculating a feature quantity of a key point detected from an object included in each of the first frame image and the second frame image; a correspondence identifying step of identifying the second frame image corresponding to each of the plurality of first frame images; A process of integrating the feature amounts calculated from the first frame images and the second frame images corresponding to each other is performed on a plurality of combinations of the first frame images and the second frame images corresponding to each other. By doing so, an integration step of creating an integrated query in which the integrated feature values are arranged in chronological order; An image processing method is provided for performing
- the computer query acquisition means for acquiring a first query moving image having a plurality of time-series first frame images and at least one second query moving image having a plurality of time-series second frame images; feature quantity calculation means for calculating a feature quantity of a key point detected from an object included in each of the first frame image and the second frame image; correspondence identifying means for identifying the second frame images corresponding to each of the plurality of first frame images; and A process of integrating the feature amounts calculated from the first frame images and the second frame images corresponding to each other is performed on a plurality of combinations of the first frame images and the second frame images corresponding to each other. Integrating means for creating an integrated query in which the integrated feature values are arranged in chronological order, A program is provided to act as a
- query acquisition means for acquiring a first query still image and at least one second query still image
- feature quantity calculation means for calculating a feature quantity of a key point detected from an object included in each of the first query still image and the second query still image
- integration means for integrating the feature amounts calculated from the first query still image and the second query still image to create an integrated query
- search means for searching still images using the integrated query as a key; has The search means is set based on the degree of similarity between each of the plurality of types of feature amounts calculated from the first query still image and each of the plurality of types of feature amounts calculated from the second query still image.
- An image processing device is provided that searches for a still image using the weighting value for each feature amount.
- the computer a query acquisition step of acquiring a first query still image and at least one second query still image; a feature quantity calculation step of calculating a feature quantity of a key point detected from an object included in each of the first query still image and the second query still image; an integration step of integrating the feature amounts calculated from the first query still image and the second query still image to create an integrated query; a search step of searching for still images using the integrated query as a key; and run In the searching step, the similarity between each of the plurality of types of feature amounts calculated from the first query still image and each of the plurality of types of feature amounts calculated from the second query still image is set based on An image processing method is provided for retrieving a still image using the weighting value for each feature amount.
- the computer query acquisition means for acquiring a first query still image and at least one second query still image; feature quantity calculation means for calculating a feature quantity of a key point detected from an object included in each of the first query still image and the second query still image; integration means for integrating the feature amounts calculated from the first query still image and the second query still image to create an integrated query; and Search means for searching still images using the integrated query as a key; function as The search means is set based on the degree of similarity between each of the plurality of types of feature amounts calculated from the first query still image and each of the plurality of types of feature amounts calculated from the second query still image.
- a program for retrieving a still image using the weighting value for each feature amount is provided.
- FIG. 1 is a configuration diagram showing an overview of an image processing apparatus according to an embodiment
- FIG. 1 is a configuration diagram showing the configuration of an image processing apparatus according to Embodiment 1
- FIG. 4 is a flowchart showing an image processing method according to Embodiment 1
- 4 is a flowchart showing a classification method according to Embodiment 1
- 4 is a flowchart showing a search method according to Embodiment 1
- FIG. 5 is a diagram showing an example of detection of a skeletal structure according to Embodiment 1
- 1 is a diagram showing a human body model according to Embodiment 1;
- FIG. 5 is a diagram showing an example of detection of a skeletal structure according to Embodiment 1;
- FIG. 5 is a diagram showing an example of detection of a skeletal structure according to Embodiment 1;
- FIG. 5 is a diagram showing an example of detection of a skeletal structure according to Embodiment 1;
- 5 is a graph showing a specific example of a classification method according to Embodiment 1;
- FIG. 8 is a diagram showing a display example of classification results according to the first embodiment;
- FIG. 2 is a diagram for explaining a search method according to Embodiment 1;
- FIG. FIG. 2 is a diagram for explaining a search method according to Embodiment 1;
- FIG. FIG. 2 is a diagram for explaining a search method according to Embodiment 1;
- FIG. FIG. 2 is a diagram for explaining a search method according to Embodiment 1;
- FIG. FIG. 2 is a diagram for explaining a search method according to Embodiment 1;
- FIG. 2 is a diagram for explaining a search method according to Embodiment 1;
- FIG. FIG. 7 is a diagram showing a display example of search results according to Embodiment 1;
- 2 is a configuration diagram showing the configuration of an image processing apparatus according to Embodiment 2;
- FIG. 9 is a flowchart showing an image processing method according to Embodiment 2;
- 10 is a flowchart showing a specific example 1 of a height pixel number calculation method according to Embodiment 2.
- FIG. 10 is a diagram showing a human body model according to Embodiment 2;
- FIG. 10 is a diagram showing an example of detection of a skeletal structure according to Embodiment 2;
- 10 is a histogram for explaining a height pixel number calculation method according to Embodiment 2;
- FIG. 10 is a diagram showing an example of detection of a skeletal structure according to Embodiment 2;
- FIG. 10 is a diagram showing a three-dimensional human body model according to Embodiment 2;
- FIG. 11 is a diagram for explaining a height pixel number calculation method according to Embodiment 2;
- FIG. FIG. 11 is a diagram for explaining a height pixel number calculation method according to Embodiment 2;
- FIG. 11 is a diagram for explaining a height pixel number calculation method according to Embodiment 2;
- FIG. 10 is a diagram for explaining a normalization method according to Embodiment 2;
- FIG. 10 is a diagram for explaining a normalization method according to Embodiment 2;
- FIG. 10 is a diagram for explaining a normalization method according to Embodiment 2;
- FIG. It is a figure which shows the hardware structural example of an image processing apparatus.
- FIG. 11 is a configuration diagram showing the configuration of an image processing apparatus according to Embodiment 3;
- FIG. 12 is a diagram for explaining query frame selection processing according to the third embodiment;
- FIG. FIG. 12 is a diagram for explaining query frame selection processing according to the third embodiment;
- FIG. 11 is a diagram for explaining a process of calculating a direction of change in feature amount according to the third embodiment
- FIG. 10 is a flow chart showing an example of the flow of processing of the image processing apparatus according to Embodiment 3
- FIG. 11 is a configuration diagram showing the configuration of an image processing apparatus according to Embodiment 3
- 10 is a flow chart showing an example of the flow of processing of the image processing apparatus according to Embodiment 3
- FIG. 12 is a diagram for explaining the concept of processing by the image processing apparatuses according to Embodiments 4 and 5
- FIG. 12 is a diagram for explaining the concept of processing by the image processing apparatuses according to Embodiments 4 and 5
- FIG. 12 is a diagram for explaining the concept of processing by the image processing apparatuses according to Embodiments 4 and 5;
- FIG. 12 is a configuration diagram showing an example of the configuration of an image processing apparatus according to Embodiment 4;
- FIG. 12 is a configuration diagram showing an example of the configuration of an image processing apparatus according to Embodiment 4;
- FIG. 12 is a diagram showing the details of processing by the image processing apparatus according to the fourth embodiment;
- FIG. 12 is a diagram showing the details of processing by the image processing apparatus according to the fourth embodiment;
- FIG. 12 is a diagram showing the details of processing by the image processing apparatus according to the fourth embodiment;
- FIG. 13 is a flow chart showing an example of the flow of processing of an image processing apparatus according to Embodiment 4;
- 13 is a flow chart showing an example of the flow of processing of an image processing apparatus according to Embodiment 4;
- FIG. 12 is a configuration diagram showing an example of the configuration of an image processing apparatus according to Embodiment 5;
- FIG. 12 is a diagram showing the details of processing by the image processing apparatus according to Embodiment 5;
- 13 is a flow chart showing an example of the flow of processing of an image processing apparatus according to Embodiment 4;
- Non-Patent Document 1 skeleton estimation technology
- Related skeleton estimation techniques such as OpenPose disclosed in Non-Patent Document 1, estimate a person's skeleton by learning various patterns of correct-correct image data.
- OpenPose disclosed in Non-Patent Document 1
- the skeletal structure estimated by skeletal estimation techniques such as OpenPose consists of "keypoints", which are characteristic points such as joints, and "bones (bone links)", which indicate links between keypoints. .
- keypoints characteristic points such as joints
- bones bone links
- FIG. 1 shows an overview of an image processing device 10 according to an embodiment.
- the image processing device 10 includes a skeleton detection section 11 , a feature amount calculation section 12 and a recognition section 13 .
- the skeleton detection unit 11 detects two-dimensional skeleton structures of a plurality of persons based on two-dimensional images acquired from a camera or the like.
- the feature amount calculation unit 12 calculates feature amounts of a plurality of two-dimensional skeleton structures detected by the skeleton detection unit 11 .
- the recognition unit 13 performs recognition processing of states of a plurality of persons based on the similarities of the plurality of feature amounts calculated by the feature amount calculation unit 12 .
- Recognition processing includes classification processing, search processing, and the like of a person's state.
- the 2D skeletal structure of a person is detected from a 2D image, and recognition processing such as classification and retrieval of the state of the person is performed based on the feature amount calculated from the 2D skeletal structure.
- recognition processing such as classification and retrieval of the state of the person is performed based on the feature amount calculated from the 2D skeletal structure.
- FIG. 2 shows the configuration of the image processing apparatus 100 according to this embodiment.
- the image processing apparatus 100 constitutes an image processing system 1 together with a camera 200 and a database (DB) 201 .
- An image processing system 1 including an image processing apparatus 100 is a system that classifies and searches for a person's posture, action, or other state based on a person's skeletal structure estimated from an image.
- the camera 200 is an imaging unit such as a surveillance camera that generates a two-dimensional image.
- the camera 200 is installed at a predetermined location and captures an image of a person or the like in an imaging area from the installation location.
- the camera 200 is directly connected or connected via a network or the like so as to be able to output captured images (video) to the image processing apparatus 100 .
- the camera 200 may be provided inside the image processing apparatus 100 .
- the database 201 is a database that stores information (data) necessary for processing of the image processing apparatus 100, processing results, and the like.
- the database 201 contains images acquired by the image acquisition unit 101, detection results of the skeletal structure detection unit 102, data for machine learning, feature amounts calculated by the feature amount calculation unit 103, classification results of the classification unit 104, search unit 105 store search results, etc.
- the database 201 is directly connected to the image processing apparatus 100 so that data can be input/output as needed, or connected via a network or the like.
- the database 201 may be provided inside the image processing apparatus 100 as a nonvolatile memory such as a flash memory, a hard disk device, or the like.
- the image processing apparatus 100 includes an image acquisition unit 101, a skeleton structure detection unit 102, a feature quantity calculation unit 103, a classification unit 104, a search unit 105, an input unit 106, and a display unit 107.
- the configuration of each unit (block) is an example, and may be configured by other units as long as the method (operation) described later is possible.
- the image processing apparatus 100 is realized by a computer device such as a personal computer or a server that executes programs, for example, but may be realized by one device or by a plurality of devices on a network. good.
- the input unit 106, the display unit 107, and the like may be external devices.
- both the classification unit 104 and the search unit 105 may be provided, or only one of them may be provided.
- Both or one of the classification unit 104 and the retrieval unit 105 is a recognition unit that performs recognition processing of the person's state.
- the image acquisition unit 101 acquires a two-dimensional image including a person captured by the camera 200 .
- the image acquisition unit 101 acquires, for example, an image including a person (video including a plurality of images) captured by the camera 200 during a predetermined monitoring period. It should be noted that an image including a person prepared in advance may be acquired from the database 201 or the like instead of being acquired from the camera 200 .
- the skeletal structure detection unit 102 detects the 2D skeletal structure of the person in the image based on the acquired 2D image.
- the skeletal structure detection unit 102 detects skeletal structures of all persons recognized in the acquired image.
- the skeletal structure detection unit 102 detects the skeletal structure of a person based on recognized features such as the joints of the person, using a skeletal structure estimation technique using machine learning.
- the skeleton structure detection unit 102 uses, for example, a skeleton estimation technique such as OpenPose described in Non-Patent Document 1.
- the feature amount calculation unit 103 calculates the feature amount of the detected two-dimensional skeletal structure, associates the calculated feature amount with the image to be processed, and stores it in the database 201 .
- the feature amount of the skeletal structure indicates the features of the skeletal structure of the person, and serves as an element for classifying and retrieving the state of the person based on the skeletal structure of the person.
- this feature quantity includes a plurality of parameters (for example, classification elements to be described later).
- the feature quantity may be the feature quantity of the entire skeletal structure, the feature quantity of a part of the skeletal structure, or may include a plurality of feature quantities such as each part of the skeletal structure.
- the feature amount is a feature amount obtained by machine-learning the skeletal structure, the size of the skeletal structure from the head to the foot on the image, and the like.
- the size of the skeletal structure is the vertical height, area, etc. of the skeletal region containing the skeletal structure on the image.
- the vertical direction (height direction or vertical direction) is the vertical direction (Y-axis direction) in the image, for example, the direction perpendicular to the ground (reference plane).
- the left-right direction (horizontal direction) is the left-right direction (X-axis direction) in the image, for example, the direction parallel to the ground.
- features that are robust to classification and search processing it is preferable to use features that are robust to classification and search processing.
- a feature quantity that is robust to the person's orientation or body shape may be used.
- the classification unit 104 classifies (clusters) a plurality of skeletal structures stored in the database 201 based on the degree of similarity of feature amounts of the skeletal structures. It can be said that the classification unit 104 classifies the states of a plurality of persons based on the feature amount of the skeletal structure as the process of recognizing the states of the persons.
- the degree of similarity is the distance between features of the skeleton structure.
- the classification unit 104 may classify the skeletal structure according to the similarity of the feature amount of the entire skeletal structure, or may classify the skeletal structure according to the similarity of the feature amount of part of the skeletal structure. Both hands) and the second part (both feet, for example) may be classified according to the similarity of feature amounts.
- the posture of the person may be classified based on the feature amount of the skeletal structure of the person in each image, or the behavior of the person may be classified based on the change in the feature amount of the skeletal structure of the person in a plurality of consecutive images in time series. can be classified. That is, the classification unit 104 can classify the state of the person, including the posture and behavior of the person, based on the feature amount of the skeletal structure. For example, the classification unit 104 classifies a plurality of skeletal structures in a plurality of images captured during a predetermined monitoring period. The classification unit 104 obtains the degree of similarity between the feature quantities to be classified, and classifies the skeletal structures with a high degree of similarity into the same cluster (group of similar postures). It should be noted that the user may be allowed to specify the classification condition as in the search. The classification unit 104 stores the classification result of the skeletal structure in the database 201 and displays it on the display unit 107 .
- the search unit 105 searches a plurality of skeleton structures stored in the database 201 for a skeleton structure with a high degree of similarity to the feature quantity of the search query (query state). It can be said that the search unit 105 searches for a person's state corresponding to a search condition (query state) from among a plurality of persons' states based on the feature amount of the skeletal structure as the recognition processing of the person's state. Similar to classification, similarity is the distance between skeletal structure features.
- the search unit 105 may search based on the similarity of the feature amount of the entire skeletal structure, or may search based on the similarity of the feature amount of a part of the skeletal structure.
- Both hands) and the second part may be retrieved based on the similarity of feature amounts.
- the posture of a person may be retrieved based on the feature amount of the skeletal structure of the person in each image, or the behavior of the person may be searched based on changes in the feature amount of the skeletal structure of the person in a plurality of images that are consecutive in time series. can be searched. That is, the search unit 105 can search for a person's state, including the person's posture and behavior, based on the feature amount of the skeletal structure. For example, the search unit 105 searches feature amounts of a plurality of skeletal structures in a plurality of images captured during a predetermined monitoring period, similarly to the classification target.
- the skeleton structure (posture) specified by the user from among the classification results displayed by the classification unit 104 is used as a search query (search key).
- search query may be selected from among a plurality of unclassified skeletal structures, and the user may input a skeletal structure to be the search query.
- the search unit 105 searches for a feature amount having a high degree of similarity with the feature amount of the skeleton structure of the search query from among the feature amounts to be searched.
- the search unit 105 stores the search result of the feature amount in the database 201 and displays it on the display unit 107 .
- the input unit 106 is an input interface that acquires information input by the user who operates the image processing apparatus 100 .
- the user is a surveillance person who monitors a person in a suspicious state from an image of a surveillance camera.
- the input unit 106 is, for example, a GUI (Graphical User Interface), and receives information according to user operations from an input device such as a keyboard, mouse, or touch panel.
- the input unit 106 receives, as a search query, the skeletal structure of a specified person from among the skeletal structures (postures) classified by the classifying unit 104 .
- the display unit 107 is a display unit that displays the result of the operation (processing) of the image processing apparatus 100, and is, for example, a display device such as a liquid crystal display or an organic EL (Electro Luminescence) display.
- a display unit 107 displays the classification result of the classification unit 104 and the search result of the search unit 105 on a GUI according to the degree of similarity or the like.
- FIG. 39 is a diagram showing a hardware configuration example of the image processing apparatus 100.
- the image processing apparatus 100 has a bus 1010 , a processor 1020 , a memory 1030 , a storage device 1040 , an input/output interface 1050 and a network interface 1060 .
- the bus 1010 is a data transmission path for the processor 1020, the memory 1030, the storage device 1040, the input/output interface 1050, and the network interface 1060 to exchange data with each other.
- the method of connecting processors 1020 and the like to each other is not limited to bus connection.
- the processor 1020 is a processor realized by a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or the like.
- the memory 1030 is a main memory implemented by RAM (Random Access Memory) or the like.
- the storage device 1040 is an auxiliary storage device realized by a HDD (Hard Disk Drive), SSD (Solid State Drive), memory card, ROM (Read Only Memory), or the like.
- the storage device 1040 stores program modules for realizing each function of the image processing apparatus 100 (for example, the image acquisition unit 101, the skeleton structure detection unit 102, the feature value calculation unit 103, the classification unit 104, the search unit 105, and the input unit 106). is doing.
- Each function corresponding to the program module is realized by the processor 1020 reading each program module into the memory 1030 and executing it.
- Storage device 1040 may also function as database 201 .
- the input/output interface 1050 is an interface for connecting the image processing apparatus 100 and various input/output devices. If the database 201 is located outside the image processing apparatus 100 , the image processing apparatus 100 may be connected to the database 201 via the input/output interface 1050 .
- a network interface 1060 is an interface for connecting the image processing apparatus 100 to a network.
- This network is, for example, a LAN (Local Area Network) or a WAN (Wide Area Network).
- a method for connecting the network interface 1060 to the network may be a wireless connection or a wired connection.
- Image processing device 100 may communicate with camera 200 via network interface 1060 . If the database 201 is located outside the image processing apparatus 100 , the image processing apparatus 100 may be connected to the database 201 via the network interface 1060 .
- An example of the hardware configuration of the image processing apparatus 100 is the same in all the following embodiments.
- FIG. 3 to 5 show the operation of the image processing apparatus 100 according to this embodiment.
- 3 shows the flow from image acquisition to search processing in the image processing apparatus 100
- FIG. 4 shows the flow of the classification processing (S104) in FIG. 3
- FIG. 5 shows the search processing (S105) in FIG. showing the flow.
- the image processing device 100 acquires an image from the camera 200 (S101).
- the image acquisition unit 101 acquires an image of a person in order to classify or search from the skeletal structure, and stores the acquired image in the database 201 .
- the image acquisition unit 101 acquires, for example, a plurality of images captured during a predetermined monitoring period, and performs subsequent processing on all persons included in the plurality of images.
- FIG. 6 shows an example of skeletal structure detection. As shown in FIG. 6, an image acquired from a monitoring camera or the like includes a plurality of persons, and the skeletal structure of each person included in the image is detected.
- FIG. 7 shows the skeletal structure of the human body model 300 detected at this time
- FIGS. 8 to 10 show detection examples of the skeletal structure.
- a skeleton structure detection unit 102 detects the skeleton structure of a human body model (two-dimensional skeleton model) 300 as shown in FIG.
- the human body model 300 is a two-dimensional model composed of key points such as human joints and bones connecting the key points.
- the skeletal structure detection unit 102 extracts feature points that can be keypoints from the image, refers to information obtained by machine learning the image of the keypoints, and detects each keypoint of the person.
- the key points of the person are head A1, neck A2, right shoulder A31, left shoulder A32, right elbow A41, left elbow A42, right hand A51, left hand A52, right hip A61, left hip A62, right knee A71. , left knee A72, right foot A81, and left foot A82.
- B72 is detected.
- the skeletal structure detection unit 102 stores the detected skeletal structure of the person in the database 201 .
- FIG. 8 is an example of detecting a person standing upright.
- an upright person is imaged from the front, and bones B1, B51 and B52, B61 and B62, and B71 and B72 viewed from the front are detected without overlapping each other.
- the bones B61 and B71 are slightly more bent than the left leg bones B62 and B72.
- Fig. 9 is an example of detecting a person who is crouching.
- a crouching person is imaged from the right side, and bones B1, B51 and B52, B61 and B62, and B71 and B72 are detected from the right side, and the right leg bone B61 is detected. And the bone B71 and the bones B62 and B72 of the left leg are greatly bent and overlapped.
- FIG. 10 is an example of detecting a sleeping person.
- a sleeping person is imaged obliquely from the front left, bones B1, B51 and B52, bones B61 and B62, bones B71 and B72 are detected from the oblique front left, and bones B71 and B72 are detected.
- the bones B61 and B71 of the left leg and the bones B62 and B72 of the left leg are bent and overlapped.
- the image processing apparatus 100 calculates the feature quantity of the detected skeletal structure (S103). For example, when the height and area of a skeletal region are used as feature quantities, the feature quantity calculator 103 extracts a region containing a skeletal structure and obtains the height (number of pixels) and area (pixel area) of that region. The height and area of the skeletal region are obtained from the coordinates of the edge of the extracted skeletal region and the coordinates of the keypoints of the edge. The feature amount calculation unit 103 stores the obtained feature amount of the skeletal structure in the database 201 .
- a skeletal region including all bones is extracted from the skeletal structure of an upright person.
- the upper end of the skeletal region is the head key point A1
- the lower end of the skeletal region is the left leg key point A82
- the left end of the skeletal region is the right elbow key point A41
- the right end of the skeletal region is the left hand key point A52.
- the height of the skeletal region is obtained from the difference between the Y coordinates of the keypoint A1 and the keypoint A82.
- the width of the skeleton region is obtained from the difference between the X coordinates of the key points A41 and A52, and the area is obtained from the height and width of the skeleton region.
- a skeletal region including all bones is extracted from the skeletal structure of a squatting person.
- the upper end of the skeletal region is the head key point A1
- the lower end of the skeletal region is the right leg key point A81
- the left end of the skeletal region is the right hip key point A61
- the right end of the skeletal region is the right hand key point A51.
- the height of the skeletal region is obtained from the difference between the Y coordinates of the keypoints A1 and A81.
- the width of the skeleton region is obtained from the difference between the X coordinates of the key points A61 and A51, and the area is obtained from the height and width of the skeleton region.
- a skeletal region including all bones is extracted from the skeletal structure of a person lying down in the horizontal direction of the image.
- the upper end of the skeletal region is the left shoulder key point A32
- the lower end of the skeletal region is the left hand key point A52
- the left end of the skeletal region is the right hand key point A51
- the right end of the skeletal region is the left foot key point A82. Therefore, the height of the skeletal region is obtained from the difference between the Y coordinates of the keypoints A32 and A52.
- the width of the skeleton region is obtained from the difference between the X coordinates of the key points A51 and A82, and the area is obtained from the height and width of the skeleton region.
- the image processing apparatus 100 performs classification processing (S104).
- the classification unit 104 calculates the similarity of the calculated feature amount of the skeletal structure (S111), and classifies the skeletal structure based on the calculated feature amount (S112). .
- the classification unit 104 obtains the degree of similarity of feature amounts between all the skeletal structures stored in the database 201 to be classified, and classifies (clusters) the skeletal structures (postures) with the highest degree of similarity into the same cluster. . Furthermore, the similarity between the classified clusters is calculated and classified, and the classification is repeated until a predetermined number of clusters is obtained.
- FIG. 11 shows an image of the classification result of the feature amount of the skeletal structure.
- FIG. 11 is an image of cluster analysis using two-dimensional classification elements.
- the two classification elements are, for example, the height of the skeleton region and the area of the skeleton region.
- a plurality of skeletal structure feature quantities are classified into three clusters C1 to C3.
- Clusters C1 to C3 correspond to postures such as a standing posture, a sitting posture, and a lying posture, and skeletal structures (persons) are classified for each similar posture.
- various classification methods can be used by classifying based on the feature amount of the skeletal structure of a person.
- the classification method may be set in advance, or may be arbitrarily set by the user.
- the classification may be performed by the same method as the retrieval method described later. In other words, classification may be performed using classification conditions similar to the search conditions.
- the classification unit 104 classifies according to the following classification method. Any classification method may be used, or a combination of arbitrarily selected classification methods may be used.
- Classification method 1 Classification according to a plurality of hierarchies Classification is performed by hierarchically combining classification according to the skeletal structure of the whole body, classification according to the skeletal structure of the upper and lower bodies, classification according to the skeletal structure of the arms and legs, and the like. That is, classification may be performed based on the feature amounts of the first portion and the second portion of the skeletal structure, and further, the feature amounts of the first portion and the second portion may be weighted for classification.
- Classification method 2 Classification based on a plurality of images in time series Classification is performed based on the feature amount of the skeletal structure in a plurality of images that are consecutive in time series. For example, feature amounts may be accumulated in the time-series direction and classified based on the cumulative value. Further, the classification may be based on the change (variation amount) of the feature amount of the skeletal structure in a plurality of consecutive images.
- the classification unit 104 displays the classification result of the skeletal structure (S113).
- the classification unit 104 acquires necessary skeletal structures and images of persons from the database 201, and displays the skeletal structures and persons for each similar posture (cluster) on the display unit 107 as a classification result.
- FIG. 12 shows a display example when postures are classified into three. For example, as shown in FIG. 12, posture areas WA1 to WA3 for each posture are displayed in the display window W1, and the skeletal structure and the person (image) of the posture respectively corresponding to the posture areas WA1 to WA3 are displayed.
- the posture area WA1 is, for example, a display area for a standing posture, and displays a skeletal structure and a person that are classified into the cluster C1 and resemble a standing posture.
- the posture area WA2 is, for example, a display area for a sitting posture, and displays a skeletal structure and a person that are classified into the cluster C2 and resemble a sitting posture.
- the posture area WA3 is, for example, a display area of a sleeping posture, and displays a skeletal structure and a person that are classified into the cluster C2 and resemble a sleeping posture.
- the image processing apparatus 100 performs search processing (S105).
- the search unit 105 receives input of search conditions (S121), and searches for a skeletal structure based on the search conditions (S122).
- the search unit 105 receives an input of a search query, which is a search condition, from the input unit 106 according to user's operation.
- the user designates (selects) the skeletal structure of the posture to be searched from among the posture areas WA1 to WA3 displayed in the display window W1. .
- the search unit 105 uses the skeleton structure specified by the user as a search query, and searches for skeleton structures with high similarity in feature quantity from among all skeleton structures stored in the database 201 to be searched.
- the search unit 105 calculates the degree of similarity between the feature quantity of the skeleton structure of the search query and the feature quantity of the skeleton structure to be searched, and extracts the skeleton structure with the calculated degree of similarity higher than a predetermined threshold.
- a feature amount calculated in advance may be used, or a feature amount obtained at the time of searching may be used.
- the search query may be input by moving each part of the skeletal structure according to the user's operation, or may be a posture demonstrated by the user in front of the camera as the search query.
- search unit 105 searches using the following search method. Any search method may be used, or any combination of search methods may be selected.
- search may be performed by combining multiple search methods (search conditions) with logical expressions (for example, AND (logical product), OR (logical sum), and NOT (negative)).
- search condition may be "(posture raising right hand) AND (posture raising left leg)".
- search method 2 When a part of the person's body is hidden in the partial search image, the search is performed using only the information of the recognizable part. For example, as in the skeletal structures 511 and 512 in FIG. 14, even if the keypoint of the left foot cannot be detected because the left foot is hidden, it can be searched using the feature amounts of other detected keypoints. Therefore, it can be determined that the skeletal structures 511 and 512 have the same posture at the time of retrieval (at the time of classification). In other words, classification and retrieval can be performed using feature amounts of some keypoints instead of all keypoints. In the example of the skeletal structures 521 and 522 in FIG.
- the feature amount of the upper body key points (A1, A2, A31, A32, A41, A42, A51, A52) is used as the search query. Therefore, it can be determined that they are in the same posture. Also, a portion (feature point) to be searched may be searched with a weight, or the threshold for similarity determination may be changed. When a part of the body is hidden, the hidden part may be ignored, or the hidden part may be taken into account in the search. By searching including hidden parts, it is possible to search postures in which the same part is hidden.
- (Search method 3) Search ignoring the left and right of the skeletal structure
- the skeletal structure of the person whose right and left sides are opposite to each other is searched as the same skeletal structure.
- the skeletal structure 531 and the skeletal structure 532 differ in the positions of the right hand key point A51, the right elbow key point A41, the left hand key point A52, and the left elbow key point A42. are the same.
- Search method 4 Search using vertical and horizontal feature amounts After performing a search using only the person's vertical (Y-axis) feature amount, the obtained result is further used for the person's horizontal (X-axis) feature amount. to search.
- search method 5 Search using a plurality of images in time series A search is performed based on the feature amount of the skeletal structure in a plurality of images that are consecutive in time series. For example, feature amounts may be accumulated in the time-series direction and searched based on the cumulative value. Furthermore, the search may be performed based on the change (variation amount) of the feature amount of the skeletal structure in a plurality of consecutive images.
- the search unit 105 displays the search result of the skeletal structure (S123).
- the search unit 105 acquires necessary skeletal structures and images of persons from the database 201 and displays the skeletal structures and persons obtained as search results on the display unit 107 .
- search results are displayed for each search query.
- FIG. 17 shows a display example when searching by three search queries (postures).
- the skeletal structures and persons of the specified search queries Q10, Q20, and Q30 are displayed on the left end, and each search query is displayed on the right side of the search queries Q10, Q20, and Q30.
- the skeletal structures and persons of the search results Q11, Q21, and Q31 are displayed side by side.
- the order in which the search results are displayed next to the search query may be the order in which the relevant skeletal structure was found, or the order in which the similarity is high.
- the parts (feature points) of the partial search are weighted for retrieval, they may be displayed in the order of similarity calculated by weighting. It may be displayed in the order of similarity calculated only from the portions (feature points) selected by the user. Further, images (frames) before and after the time series may be cut out for a certain period of time and displayed, centering on the image (frame) of the search result.
- the present embodiment it is possible to detect the skeletal structure of a person from a two-dimensional image and perform classification and retrieval based on the feature amount of the detected skeletal structure. As a result, it is possible to classify the postures having a high degree of similarity, and to search for similar postures having a high degree of similarity with the search query (search key).
- search key By classifying and displaying similar postures from the image, the posture of the person in the image can be grasped without the user specifying the posture or the like. Since the user can specify the posture of the search query from among the classification results, even if the user does not know the details of the posture to be searched in advance, it is possible to search for the desired posture.
- classification and retrieval can be performed using the whole or part of a person's skeletal structure as a condition, enabling flexible classification and retrieval.
- Embodiment 2 will be described below with reference to the drawings.
- the feature amount is obtained by normalization using the height of the person. Others are the same as those of the first embodiment.
- FIG. 18 shows the configuration of the image processing apparatus 100 according to this embodiment.
- the image processing apparatus 100 further includes a height calculator 108 in addition to the configuration of the first embodiment. Note that the feature amount calculation unit 103 and the height calculation unit 108 may be integrated into one processing unit.
- a height calculation unit (height estimation unit) 108 calculates the height of a person in a two-dimensional image when standing upright (height in pixels) based on the two-dimensional skeletal structure detected by the skeletal structure detection unit 102 ( presume. It can also be said that the number of height pixels is the height of the person in the two-dimensional image (the length of the whole body of the person in the two-dimensional image space). The height calculation unit 108 obtains the number of height pixels (the number of pixels) from the length of each bone of the detected skeletal structure (the length in the two-dimensional image space).
- specific examples 1 to 3 are used as the method for obtaining the height pixel count. Any one of the methods of Examples 1 to 3 may be used, or a plurality of arbitrarily selected methods may be used in combination.
- the number of height pixels is obtained by totaling the length of the bones from the head to the feet among the bones of the skeletal structure. If the skeletal structure detection unit 102 (skeletal structure estimation technology) does not output the top of the head and the feet, it can be corrected by multiplying by a constant as necessary.
- the number of height pixels is calculated using a human body model that indicates the relationship between the length of each bone and the length of the whole body (height in a two-dimensional image space).
- the number of height pixels is calculated by fitting a three-dimensional human body model to a two-dimensional skeletal structure.
- the feature amount calculation unit 103 of the present embodiment is a normalization unit that normalizes the skeletal structure (skeletal information) of the person based on the calculated number of pixels of the height of the person.
- the feature amount calculation unit 103 stores the normalized feature amount (normalized value) of the skeletal structure in the database 201 .
- the feature amount calculation unit 103 normalizes the height on the image of each key point (feature point) included in the skeletal structure by the number of height pixels.
- the height direction is the vertical direction (Y-axis direction) in the two-dimensional coordinate (XY coordinate) space of the image.
- the height of the keypoint can be obtained from the Y coordinate value (the number of pixels) of the keypoint.
- the height direction may be the direction of the vertical projection axis (vertical projection direction) obtained by projecting the direction of the vertical axis perpendicular to the ground (reference plane) in the three-dimensional coordinate space of the real world onto the two-dimensional coordinate space.
- the height of the keypoint is obtained by calculating the vertical projection axis by projecting the axis perpendicular to the ground in the real world onto the two-dimensional coordinate space based on the camera parameters, and calculating the value along this vertical projection axis (the number of pixels ) can be obtained from
- the camera parameters are imaging parameters of an image.
- the camera parameters are the attitude, position, imaging angle, focal length, etc. of the camera 200 .
- the camera 200 an object whose length and position are known in advance can be imaged, and camera parameters can be obtained from the image. Distortion occurs at both ends of the captured image, and the vertical direction of the real world may not match the vertical direction of the image.
- the parameters of the camera that captured the image it is possible to know how much the vertical direction in the real world is tilted in the image. Therefore, by normalizing the values of the keypoints along the vertical projection axis projected into the image based on the camera parameters by the height, it is possible to convert the keypoints into features considering the deviation between the real world and the image. can.
- the left-right direction is the left-right direction (X-axis direction) in the two-dimensional coordinate (XY coordinate) space of the image, or the direction parallel to the ground in the three-dimensional coordinate space of the real world. is projected onto the two-dimensional coordinate space.
- FIG. 19 to 23 show the operation of the image processing apparatus 100 according to this embodiment.
- FIG. 19 shows the flow from image acquisition to search processing in the image processing apparatus 100
- FIGS. 23 shows the flow of the normalization process (S202) in FIG.
- the image processing apparatus 100 After image acquisition (S101) and skeletal structure detection (S102), the image processing apparatus 100 performs height pixel count calculation processing based on the detected skeletal structure (S201).
- the height of the skeletal structure of the person in the image when standing upright is the number of height pixels (h)
- the height of each keypoint of the skeletal structure in the state of the person in the image is the keypoint. Let the height be (yi). Specific examples 1 to 3 of the height pixel number calculation process will be described below.
- the length of the bone from the head to the foot is used to obtain the number of pixels of the height.
- the height calculation unit 108 acquires the length of each bone (S211), and totals the acquired lengths of each bone (S212).
- the height calculation unit 108 obtains the length of the bones on the two-dimensional image from the head to the feet of the person and obtains the number of pixels of the height. 24, bone B1 (length L1), bone B51 (length L21), bone B61 (length L31) and bone B71 (length L41), or , bone B1 (length L1), bone B52 (length L22), bone B62 (length L32), and bone B72 (length L42).
- the length of each bone can be obtained from the coordinates of each keypoint in the two-dimensional image.
- the height pixel number (h) is calculated by multiplying L1+L21+L31+L41 or L1+L22+L32+L42 by a correction constant.
- the longer value is used as the number of height pixels. That is, each bone has the longest length in the image when the image is taken from the front, and is displayed to be short when the bone is tilted in the depth direction with respect to the camera. Therefore, the longer bones are more likely to be imaged from the front, and are considered to be closer to the true values. Therefore, it is preferable to choose the longer value.
- bone B1, bone B51 and bone B52, bone B61 and bone B62, bone B71 and bone B72 are detected without overlapping each other.
- the sums of these bones, L1+L21+L31+L41 and L1+L22+L32+L42, are calculated, and the value obtained by multiplying L1+L22+L32+L42 on the left leg side where the length of the detected bone is longer by a correction constant is taken as the number of height pixels.
- bone B1, bone B51 and bone B52, bone B61 and bone B62, bone B71 and bone B72 are respectively detected, and bone B61 and bone B71 of the right leg and bone B62 and bone B72 of the left leg are overlapped.
- the sums of these bones, L1+L21+L31+L41 and L1+L22+L32+L42, are calculated, and the value obtained by multiplying L1+L21+L31+L41 on the right leg side where the length of the detected bone is longer by a correction constant is taken as the height pixel number.
- bone B1, bone B51 and bone B52, bone B61 and bone B62, bone B71 and bone B72 are respectively detected, and bone B61 and bone B71 of the right leg and bone B62 and bone B72 of the left leg are overlapped.
- the sums of these bones, L1+L21+L31+L41 and L1+L22+L32+L42, are calculated, and the value obtained by multiplying L1+L22+L32+L42 on the left leg side where the length of the detected bone is longer by a correction constant is taken as the number of height pixels.
- the height can be obtained by totaling the length of the bones from the head to the feet, the number of pixels of the height can be obtained by a simple method.
- the height pixel count can be accurately calculated even when the whole person is not shown in the image, such as when the person is crouching. can be estimated.
- the number of height pixels is obtained using a two-dimensional skeleton model that indicates the relationship between the length of bones included in the two-dimensional skeleton structure and the length of the whole body of a person in the two-dimensional image space.
- FIG. 28 is a human body model (two-dimensional skeleton model) 301 used in Specific Example 2, showing the relationship between the length of each bone on the two-dimensional image space and the length of the whole body on the two-dimensional image space.
- the relationship between the length of each bone of an average person and the length of the whole body is associated with each bone of the human body model 301 .
- the length of the head bone B1 is the length of the whole body x 0.2 (20%)
- the length of the right hand bone B41 is the length of the whole body x 0.15 (15%)
- the length of the right leg is
- the length of bone B71 is the length of the whole body ⁇ 0.25 (25%).
- the average length of the whole body can be obtained from the length of each bone.
- a human body model may be prepared for each person's attributes such as age, sex, and nationality. As a result, the length of the whole body (height) can be obtained appropriately according to the attributes of the person.
- the height calculation unit 108 acquires the length of each bone (S221).
- the height calculator 108 acquires the lengths of all bones (lengths in the two-dimensional image space) in the detected skeletal structure.
- FIG. 29 shows an example in which a skeletal structure is detected by capturing an image of a squatting person from the right rear oblique direction.
- the bones of the head, left arm, and left hand cannot be detected. Therefore, the lengths of the detected bones B21, B22, B31, B41, B51, B52, B61, B62, B71, and B72 are acquired.
- the height calculation unit 108 calculates the number of height pixels from the length of each bone based on the human body model, as shown in FIG. 21 (S222).
- the height calculator 108 refers to a human body model 301 showing the relationship between each bone and the length of the whole body as shown in FIG. 28, and obtains the number of height pixels from the length of each bone.
- the length of the bone B41 on the right hand is the length of the whole body ⁇ 0.15
- the length of the bone B41/0.15 is used to obtain the height pixel number based on the bone B41.
- the length of the bone B71 of the right leg is the length of the whole body ⁇ 0.25, the length of the bone B71/0.25 is used to obtain the height pixel number based on the bone B71.
- the human body model referred to at this time is, for example, the human body model of an average person, but the human body model may be selected according to the attributes of the person, such as age, gender, and nationality. For example, when a person's face is shown in the captured image, the person's attribute is identified based on the face, and a human body model corresponding to the identified attribute is referred to. By referring to machine-learned face information for each attribute, it is possible to recognize a person's attribute from the facial features in the image. Also, when the attributes of a person cannot be identified from the image, an average human body model may be used.
- the height pixel count calculated from the length of the bone may be corrected by camera parameters. For example, if the camera is placed at a high position and the person is shot looking down, the horizontal length of the shoulder bones, etc. in the two-dimensional skeletal structure is not affected by the camera's depression angle, but the vertical length of the neck-waist bones, etc. The length decreases as the depression angle of the camera increases. As a result, the height pixel count calculated from the horizontal length of the shoulder-width bone tends to be larger than the actual number. Therefore, by using the camera parameters, it is possible to know at what angle the person is looking down at the camera. This makes it possible to more accurately calculate the number of height pixels.
- the height calculation unit 108 calculates the optimal value of the number of height pixels, as shown in FIG. 21 (S223).
- the height calculation unit 108 calculates the optimal value of the height pixel count from the height pixel count obtained for each bone. For example, as shown in FIG. 30, a histogram of the number of height pixels obtained for each bone is generated, and the largest number of height pixels is selected. In other words, among the plurality of height pixel numbers obtained based on the plurality of bones, the height pixel number that is longer than the others is selected. For example, the upper 30% are set as effective values, and the number of height pixels by bones B71, B61, and B51 in FIG. 30 is selected.
- the average of the selected height pixel counts may be obtained as the optimum value, or the maximum height pixel count may be obtained as the optimum value. Since the height is calculated from the length of the bones in the two-dimensional image, if the bones are not formed from the front, that is, if the bones are photographed tilted in the depth direction when viewed from the camera, the length of the bones will be measured from the front. shorter than the case. Then, a value with a large height pixel count is more likely to have been captured from the front than a value with a small height pixel count, and is a more plausible value.
- a human body model showing the relationship between bones in a two-dimensional image space and the length of the whole body is used to obtain the number of height pixels based on the bones of the detected skeletal structure. Even if is not obtained, the number of height pixels can be obtained from some bones. In particular, by adopting the larger value among the values obtained from a plurality of bones, the number of height pixels can be estimated with high accuracy.
- a 2D skeletal structure is fitted to a 3D human body model (3D skeletal model), and a whole body skeletal vector is obtained using the number of height pixels of the fitted 3D human body model.
- the height calculation unit 108 first calculates camera parameters based on the image captured by the camera 200 (S231).
- the height calculator 108 extracts an object whose length is known in advance from a plurality of images captured by the camera 200, and obtains camera parameters from the size (number of pixels) of the extracted object. Note that the camera parameters may be obtained in advance, and the obtained camera parameters may be obtained as necessary.
- the height calculation unit 108 adjusts the placement and height of the 3D human body model (S232).
- the height calculation unit 108 prepares a three-dimensional human body model for height pixel number calculation for the detected two-dimensional skeletal structure, and arranges it in the same two-dimensional image based on the camera parameters.
- the "relative positional relationship between the camera and the person in the real world" is specified from the camera parameters and the two-dimensional skeleton structure. For example, if the position of the camera is assumed to be coordinates (0, 0, 0), the coordinates (x, y, z) of the person's standing (or sitting) position are specified. Then, by assuming an image when the 3D human body model is arranged at the same position (x, y, z) as the specified person and captured, the 2D skeletal structure and the 3D human body model are superimposed.
- FIG. 31 is an example of detecting a two-dimensional skeletal structure 401 by capturing an image of a crouching person from the front left diagonally.
- the two-dimensional skeleton structure 401 has two-dimensional coordinate information. It is preferable that all bones are detected, but some bones may not be detected.
- a three-dimensional human body model 402 as shown in FIG. 32 is prepared for this two-dimensional skeletal structure 401 .
- a three-dimensional human body model (three-dimensional skeleton model) 402 is a skeleton model having three-dimensional coordinate information and having the same shape as the two-dimensional skeleton structure 401 .
- a prepared three-dimensional human body model 402 is arranged and superimposed on the detected two-dimensional skeletal structure 401 . Also, the height of the three-dimensional human body model 402 is adjusted so as to match the two-dimensional skeletal structure 401 while being superimposed.
- the three-dimensional human body model 402 prepared at this time may be a model in a state close to the posture of the two-dimensional skeletal structure 401 as shown in FIG. 33, or may be a model in an upright state.
- a technique of estimating a posture in a three-dimensional space from a two-dimensional image using machine learning may be used to generate the three-dimensional human body model 402 with the estimated posture.
- a three-dimensional posture can be estimated from a two-dimensional image by learning joint information in a two-dimensional image and joints in a three-dimensional space.
- the height calculation unit 108 fits the 3D human body model to the 2D skeletal structure as shown in FIG. 22 (S233). As shown in FIG. 34, the height calculation unit 108 calculates the three-dimensional human body model 402 so that the postures of the three-dimensional human body model 402 and the two-dimensional skeletal structure 401 match each other in a state in which the three-dimensional human body model 402 is superimposed on the two-dimensional skeletal structure 401 .
- the dimensional human body model 402 is deformed. That is, the height, body orientation, and joint angles of the three-dimensional human body model 402 are adjusted so that the difference from the two-dimensional skeletal structure 401 is optimized.
- the joints of the three-dimensional human body model 402 are rotated within the human movable range, the entire three-dimensional human body model 402 is rotated, and the overall size is adjusted.
- the fitting between the three-dimensional human body model and the two-dimensional skeletal structure is performed in a two-dimensional space (two-dimensional coordinates). That is, the three-dimensional human body model is mapped into a two-dimensional space, and the three-dimensional human body model is transformed into a two-dimensional skeletal structure in consideration of how the deformed three-dimensional human body model changes in the two-dimensional space (image). Optimize.
- the height calculation unit 108 calculates the number of height pixels of the fitted three-dimensional human body model, as shown in FIG. 22 (S234).
- the difference between the three-dimensional human body model 402 and the two-dimensional skeletal structure 401 disappears and the postures match as shown in FIG.
- the height pixel number is calculated from the bone length (pixel number) from the head to the feet when the three-dimensional human body model 402 is erected.
- the lengths of the bones from the head to the feet of the three-dimensional human body model 402 may be totaled.
- the 3D human body model is fitted to the 2D skeletal structure based on the camera parameters, and the number of height pixels is obtained based on the 3D human body model. That is, even if the error is large because all the bones are projected obliquely, the number of height pixels can be estimated with high accuracy.
- the image processing apparatus 100 performs normalization processing (S202) following the height pixel count calculation processing.
- the feature amount calculation unit 103 calculates the keypoint height (S241).
- the feature amount calculation unit 103 calculates the keypoint heights (the number of pixels) of all keypoints included in the detected skeletal structure.
- the keypoint height is the length (number of pixels) in the height direction from the lowest end of the skeletal structure (for example, the keypoint of one of the legs) to that keypoint.
- the keypoint height is obtained from the Y coordinate of the keypoint in the image.
- the keypoint height may be obtained from the length in the direction along the vertical projection axis based on the camera parameters.
- the height (yi) of the neck keypoint A2 is the Y coordinate of the keypoint A2 minus the Y coordinate of the right leg keypoint A81 or the left leg keypoint A82.
- a reference point is a reference point for representing the relative height of a keypoint.
- the reference point may be set in advance or may be selected by the user.
- the reference point is preferably the center of the skeletal structure or higher than the center (above in the vertical direction of the image), for example, the coordinates of the neck key point. Note that the coordinates of the head or other key points may be used as the reference point instead of the neck.
- Arbitrary coordinates for example, the center coordinates of the skeleton structure, etc. may be used as the reference point without being limited to the key point.
- the feature amount calculation unit 103 normalizes the keypoint height (yi) by the number of height pixels (S243).
- the feature amount calculation unit 103 normalizes each keypoint using the keypoint height, the reference point, and the height pixel count of each keypoint. Specifically, the feature amount calculation unit 103 normalizes the relative height of the keypoint with respect to the reference point by the number of height pixels.
- the Y coordinate of the reference point (key point of the neck) is set to (yc), and the feature amount (normalized value) is obtained using the following equation (1).
- (yi) and (yc) are converted into values in the direction along the vertical projection axis.
- the coordinates (x0, y0), (x1, y1), . is converted into an 18-dimensional feature amount as follows.
- FIG. 36 shows an example of the feature amount of each keypoint calculated by the feature amount calculation unit 103.
- the feature amount of the key point A2 is 0.0
- the feature amount of the key point A31 of the right shoulder and the key point A32 of the left shoulder at the same height as the neck are also 0.0.
- the feature value of the keypoint A1 of the head higher than the neck is -0.2.
- the right hand keypoint A51 and left hand keypoint A52 lower than the neck have a feature quantity of 0.4
- the right foot keypoint A81 and left foot keypoint A82 have a feature quantity of 0.9.
- the feature amount (normalized value) of the present embodiment indicates the feature in the height direction (Y direction) of the skeletal structure (key point), and influences the change in the lateral direction (X direction) of the skeletal structure. do not receive
- the skeletal structure of a person is detected from a two-dimensional image, and the number of height pixels (height when standing upright in the two-dimensional image space) obtained from the detected skeletal structure is used to Normalize each keypoint of the skeleton structure.
- this normalized feature amount it is possible to improve robustness when performing classification, search, and the like. That is, the feature amount of the present embodiment is not affected by changes in the horizontal direction of the person as described above, and is therefore highly robust against changes in the orientation of the person and the body shape of the person.
- a skeleton estimation technique such as OpenPose
- a skeleton estimation technique such as OpenPose
- by normalizing the key points of the skeletal structure it is possible to obtain clear and easy-to-understand feature quantities, so unlike black-box algorithms such as machine learning, users are highly satisfied with the processing results.
- Embodiment 3 will be described below with reference to the drawings. In this embodiment, a specific example of processing for searching for a moving image including a desired scene will be described.
- FIG. 40 shows an example of a functional block diagram of the image processing apparatus 100 of this embodiment.
- the image processing apparatus 100 includes a query acquisition unit 109, a query frame selection unit 112, a skeleton structure detection unit 102, a feature amount calculation unit 103, a change calculation unit 110, and a search unit 111.
- the image processing apparatus 100 may further include other functional units described in the first and second embodiments.
- the query acquisition unit 109 acquires a query video made up of a plurality of time-series first frame images. For example, the query acquisition unit 109 acquires a query moving image (moving image file) input/designated/selected by user operation.
- a query moving image moving image file
- the query frame selection unit 112 selects at least part of the plurality of first frame images as query frames. As shown in FIGS. 41 and 42, the query frame selection unit 112 can intermittently select query frames from a plurality of time-series first frame images included in the query moving image. The number of first frame images between query frames may be constant or may be random. The query frame selection unit 112 can execute, for example, any one of the following selection processes 1 to 3.
- the query frame selection unit 112 selects a query frame based on user input. That is, the user makes an input designating at least part of the plurality of first frame images as the query frame. Then, the query frame selection unit 112 selects the first frame image specified by the user as the query frame.
- the query frame selection unit 112 selects a query frame according to a predetermined rule.
- the query frame selection unit 112 selects multiple query frames from multiple first frame images at predetermined regular intervals. That is, the query frame selection unit 112 selects a query frame every M frames. Examples of M include 2 or more and 10 or less, but are not limited thereto. M may be predetermined or may be selected by the user.
- the query frame selection unit 112 selects a query frame according to a predetermined rule.
- the query frame selection unit 112 selects one query frame, and then selects between the query frame and each of the first frame images after the query frame in chronological order. Calculate the similarity of Similarity is the same concept as in the first and second embodiments. Then, the query frame selection unit 112 selects, as a new query frame, the first frame image whose similarity is equal to or lower than the reference value and whose chronological order is the earliest.
- the query frame selection unit 112 calculates the degree of similarity between the newly selected query frame and each of the first frame images following the query frame in chronological order. Then, the query frame selection unit 112 selects, as a new query frame, the first frame image whose similarity is equal to or lower than the reference value and whose chronological order is the earliest.
- the query frame selection unit 112 selects a query frame by repeating the process. According to this process, the poses of persons included in adjacent query frames are different to some extent. Therefore, it is possible to select a plurality of query frames showing a characteristic posture of a person while suppressing an increase in the number of query frames.
- the reference value may be predetermined, may be selected by the user, or may be set by other means.
- the skeletal structure detection unit 102 detects key points of a person (object) included in each of a plurality of first frame images.
- the skeletal structure detection unit 102 may subject only the query frame to the detection process, or may subject all the first frame images to the detection process.
- the configuration of the skeletal structure detection unit 102 is the same as in Embodiments 1 and 2, so detailed description thereof will be omitted here.
- the feature amount calculation unit 103 calculates the feature amount of the detected keypoint, that is, the feature amount of the detected two-dimensional skeleton structure for each first frame image.
- the feature amount calculation unit 103 may subject only the query frame to the calculation process, or may subject all the first frame images to the calculation process. Since the configuration of the feature amount calculation unit 103 is the same as that of the first and second embodiments, detailed description thereof is omitted here.
- the change calculation unit 110 calculates the direction of change of the feature amount along the time axis of the plurality of time-series first frame images.
- the change calculation unit 110 calculates, for example, the direction of change in feature quantity between adjacent query frames.
- the feature amount is the feature amount calculated by the feature amount calculation unit 103 .
- the feature quantity is the height, area, etc. of the skeleton region, and is expressed numerically.
- the direction of change of the feature amount is divided into three directions, for example, "direction of increasing numerical value", “no change in numerical value", and "direction of decreasing numerical value”. “No numerical value change” may be a case where the absolute value of the amount of change in the feature amount is 0, or a case where the absolute value of the amount of change in the feature amount is equal to or less than the threshold.
- the change calculation unit 110 can calculate time-series data indicating a time-series change in the direction of change in the feature amount.
- the time-series data is, for example, "direction of increasing numerical value” ⁇ "direction of increasing numerical value” ⁇ “direction of increasing numerical value” ⁇ “no change in numerical value” ⁇ “no change in numerical value” ⁇ “higher numerical value direction”.
- “the direction in which the numerical value increases” is expressed as "1”
- “no change in the numerical value” is expressed as "0”
- the direction in which the numerical value decreases” is expressed as "-1”.
- 111001" can be represented by a numerical string.
- the change calculation unit 110 can calculate the direction of change in the feature amount between the two images.
- the search unit 111 searches for moving images using the direction of change in the feature amount calculated by the change calculation unit 110 as a key. Specifically, the search unit 111 searches for a DB video that matches the key from videos stored in the database 201 (hereinafter referred to as DB videos).
- the search unit 111 can execute, for example, one of the following video search processes 1 and 2.
- the search unit 111 can search for DB moving images whose similarity of the time-series data is equal to or higher than a reference value.
- a method for calculating the degree of similarity of time-series data is not particularly limited, and any technique can be adopted.
- time-series data may be created in advance by the same method as described above corresponding to each DB moving image stored in the database 201 and stored in the database.
- the search unit 111 may process each DB moving image stored in the database 201 in the same manner as described above each time a search process is performed, and create the time-series data for each DB moving image.
- the search unit 111 can search for DB moving images indicating the direction of change in the feature amount.
- index data of the direction of change of the feature amount shown in each DB moving image may be created in advance corresponding to each DB moving image stored in the database 201 and stored in the database.
- the search unit 111 processes each of the DB moving images stored in the database 201 in the same manner as described above each time a search process is performed, and determines the direction of change in the feature quantity indicated in each DB moving image for each DB moving image. index data may be created.
- FIG. 1 An example of the processing flow of the image processing apparatus 100 will be described using FIG. The purpose here is to explain the flow of processing. Since the details of each process have been described above, descriptions thereof are omitted here.
- the image processing apparatus 100 acquires a query moving image composed of a plurality of time-series first frame images (S400), it selects at least part of the plurality of first frame images as a query frame (S401).
- the image processing apparatus 100 detects keypoints of objects included in each of the plurality of first frame images (S402). Note that only the query frame selected in S401 may be subject to the processing, or all first frame images may be subject to the processing.
- the image processing apparatus 100 calculates feature amounts of the detected keypoints for each of the plurality of first frame images (S403). Note that only the query frame selected in S401 may be subject to the processing, or all first frame images may be subject to the processing.
- the image processing apparatus 100 calculates the direction of change of the feature amount along the time axis of the plurality of time-series first frame images (S404).
- the image processing apparatus 100 calculates the direction of change in feature amount between adjacent query frames.
- the direction of change is divided into, for example, three directions, namely, "direction of increase in numerical value", “direction of no change in numerical value”, and "direction of decreasing numerical value”.
- the image processing device 100 can calculate time-series data indicating a time-series change in the direction of change in the feature amount.
- the image processing apparatus 100 can calculate the direction of change in feature amount between the two images.
- the image processing apparatus 100 searches for DB moving images using the change direction of the feature amount calculated in S404 as a key (S405). Specifically, the image processing apparatus 100 searches for a DB moving image that matches the key from among the DB moving images stored in the database 201 . The image processing apparatus 100 then outputs the search result.
- the output of search results can be realized by adopting any technology.
- the image processing apparatus 100 of the present embodiment can be configured to employ one or more of Modifications 1 to 7 below.
- the image processing device 100 may not have the query frame selection unit 112 .
- the change calculator 110 can calculate the direction of change in the feature amount between adjacent first frame images.
- the change calculation unit 110 can calculate time-series data indicating a time-series change in the direction of change in the feature amount.
- the change calculation unit 110 can calculate the direction of change in feature amount between the two images.
- FIG. 1 An example of the processing flow of the image processing apparatus 100 in the modification will be described using FIG.
- the purpose here is to explain the flow of processing. Since the details of each process have been described above, descriptions thereof are omitted here.
- the image processing device 100 acquires a query moving image composed of a plurality of time-series first frame images (S300). Next, the image processing apparatus 100 detects keypoints of objects included in each of the plurality of first frame images (S301). Next, the image processing apparatus 100 calculates feature amounts of the detected keypoints for each of the plurality of first frame images (S302).
- the image processing apparatus 100 calculates the direction of change of the feature amount along the time axis of the plurality of time-series first frame images (S303). Specifically, the image processing apparatus 100 calculates the direction of change in feature amount between adjacent first frame images.
- the image processing apparatus 100 searches for DB moving images using the direction of change in the feature amount calculated in S303 as a key (S304). Specifically, the image processing apparatus 100 searches for a DB moving image that matches the key from among the DB moving images stored in the database 201 . The image processing apparatus 100 then outputs the search result.
- the output of search results can be realized by adopting any technology.
- the image processing apparatus 100 detects key points of a person's body, and searches DB moving images using the direction of change as a key.
- the image processing apparatus 100 can detect a key point of an object other than a person, and search DB moving images using the direction of change as a key.
- Objects are not particularly limited, and examples thereof include animals, plants, natural products, artificial objects, and the like.
- the change calculation unit 110 can calculate the magnitude of change in the feature amount in addition to the direction of change in the feature amount.
- the change calculation unit 110 can calculate the magnitude of change in feature quantity between adjacent query frames or between adjacent first frame images.
- the magnitude of change in the feature amount can be represented, for example, by the absolute value of the difference between the numerical values indicating the feature amount.
- the magnitude of change in feature amount may be a value obtained by normalizing the absolute value.
- the change calculation unit 110 calculates a time series that further indicates a time-series change in the magnitude of change in addition to the direction of change in the feature amount. Data can be calculated.
- the change calculation unit 110 can calculate the direction and magnitude of the change in feature amount between the two images.
- the search unit 111 searches the DB video using the change direction and change magnitude calculated by the change calculation unit 110 as keys.
- the search unit 111 can search for DB videos whose similarity of the time-series data is equal to or higher than the reference value.
- a method for calculating the degree of similarity of time-series data is not particularly limited, and any technique can be adopted.
- the search unit 111 retrieves the DB animation showing the direction and magnitude of change in the feature amount. can be searched.
- the change calculation unit 110 can calculate the speed of change of the feature amount in addition to the direction of change of the feature amount.
- This modified example is effective when query frames are selected at random intervals from the first frame image as shown in FIG. 42 and the direction of change in feature amount between adjacent query frames is calculated. In this case, it is possible to search for more similar DB moving images by referring to the speed of change in the feature amount between adjacent query frames.
- the change calculation unit 110 can calculate the speed of change in feature values between adjacent query frames.
- the speed can be calculated by dividing the amount of change in the feature amount by a value indicating the amount of time between adjacent query frames (the number of frames, a value converted into time based on the frame rate, etc.).
- the magnitude of change in the feature amount can be represented, for example, by the absolute value of the difference between the numerical values indicating the feature amount.
- the magnitude of change in feature amount may be a value obtained by normalizing the absolute value.
- the change calculation unit 110 can calculate time-series data that further indicates the speed of change in addition to the direction of change of the feature amount.
- the change calculation unit 110 can calculate the direction and speed of changes in feature amounts between the two images.
- the search unit 111 searches the DB video using the change direction and change speed calculated by the change calculation unit 110 as keys.
- the search unit 111 can search for DB videos whose similarity of the time-series data is equal to or higher than the reference value.
- a method for calculating the degree of similarity of time-series data is not particularly limited, and any technique can be adopted.
- the search unit 111 can search the DB video showing the direction and speed of change in the feature amount.
- the search unit 111 has searched for a DB moving image that matches the key, but may search for a DB moving image that does not match the key. That is, the search unit 111 may search for a DB moving image whose degree of similarity to the time-series data, which is the key, is less than a reference value. Further, the search unit 111 may search for a DB moving image that does not include the direction of change of the key feature amount (which may include magnitude, speed, etc.).
- the search unit 111 may search for DB moving images that match search conditions in which a plurality of keys are connected by arbitrary logical operators.
- the search unit 111 further extracts representative images selected from the first frame images of the query video. It can be used as a key to search DB videos.
- the number of representative images may be one or plural.
- the query frame may be used as the representative image, a frame selected from the query frames by arbitrary means may be used as the representative image, or the representative image may be selected from the first frame images by other means. .
- the search unit 111 calculates the degree of similarity between the DB moving image stored in the database 201 and the query moving image calculated based on the representative image, and the result calculated by the change calculating unit 110 (the direction of change in the special amount, A DB moving image having a total similarity equal to or greater than a reference value, which is obtained by integrating the similarity with the query moving image calculated based on the size, speed, etc., can be retrieved.
- the search unit 111 can calculate the degree of similarity between each DB moving image and the query moving image based on the following criteria.
- the degree of similarity between the representative image and the frame image is calculated based on the posture of the person included in each image. The more similar the poses are, the higher the degree of similarity between the representative image and the frame image.
- the search unit 111 may calculate the degree of similarity of the feature amount of the skeletal structure described in the above embodiment, or use other well-known techniques to calculate the posture of the person. may be calculated.
- the similarity of the time-series data is the similarity between each DB video and the query video.
- the direction of change is the same as that of the query video, and the magnitude and speed of change are shown in the query video.
- the similarity of the DB moving image is increased as it is more similar to that indicated by .
- each degree of similarity may be normalized and summed up.
- each degree of similarity may be weighted. That is, the similarity based on the representative image or the value obtained by multiplying the standard value thereof by a predetermined weighting factor, and the similarity based on the result calculated by the change calculation unit 110 (the direction, magnitude, speed, etc. of change in the characteristic amount)
- a value obtained by multiplying the degree or its standard value by a predetermined weighting factor may be calculated as the integration result.
- the same effects as those of the first and second embodiments are realized. Further, according to the image processing apparatus 100 of the present embodiment, it is possible to search for moving images using the direction of change in the orientation of an object included in an image, the magnitude of change, the speed of change, and the like as keys. According to the image processing apparatus 100 of this embodiment, it is possible to accurately search for a moving image including a desired scene.
- the image processing apparatus 100 may configure the image processing system 1 together with the camera 200 and database 201 .
- Embodiment 4 will be described below with reference to the drawings.
- a plurality of queries showing the same posture and the same action are integrated, and a search is performed using the combined query, so that images showing the same posture and action as those indicated by the query can be obtained with high accuracy. make it searchable.
- the image of the point indicated as "query” in the figure is used as the query image, and the image of the other points is used as the image to be searched.
- the query image belongs to category A. Therefore, it is preferable to implement a search in which all images belonging to category A are included in the search results and images belonging to categories other than category A are not included in the search results.
- a threshold value is searched. If the threshold value is set too high as indicated by S1 in the figure, the inconvenience that images belonging to categories other than A are included in the search results can be suppressed, but images belonging to category A may be omitted from the search results.
- the threshold value is set too low as indicated by S2 in the figure, it is possible to prevent images belonging to the category A from being omitted from the search results, but images belonging to categories other than the category A may be included in the search results. obtain.
- FIG. 49 shows an example of a functional block diagram of the image processing apparatus 100 of this embodiment.
- the image processing apparatus 100 has a query acquisition unit 109 , a skeleton structure detection unit 102 , a feature amount calculation unit 103 , a correspondence identification unit 114 and an integration unit 113 .
- the image processing apparatus 100 may further have a search section 111.
- the image processing apparatus 100 may further include other functional units described in the first to third embodiments.
- the query acquisition unit 109 acquires the first query video and at least one second query video.
- the query acquisition unit 109 acquires a first query video and a second query video input/designated/selected by user operation.
- the first query moving image has a plurality of time-series first frame images.
- the second query moving image has a plurality of time-series second frame images.
- the first query video and the second query video show the same posture or the same action.
- the time lengths of the first query video and the second query video may be the same or different. That is, the number of first frame images and the number of second frame images may be the same or different. Also, the first query video and the second query video may include the same person or may include different persons. Note that it is desirable that the angle of view of the first query moving image and the angle of view of the second query moving image are aligned to some extent depending on the feature amount to be used.
- the skeletal structure detection unit 102 detects key points of a person (object) included in each of a plurality of first frame images. Also, the skeletal structure detection unit 102 detects key points of a person (object) included in each of the plurality of second frame images. Since the configuration of the skeletal structure detection unit 102 is the same as that of Embodiments 1 to 3, detailed description thereof will be omitted here.
- the feature amount calculation unit 103 calculates the feature amount of the detected keypoint, that is, the feature amount of the detected two-dimensional skeleton structure for each first frame image. Also, the feature amount calculation unit 103 calculates the feature amount of the detected keypoint, that is, the feature amount of the detected two-dimensional skeleton structure, for each second frame image. Since the configuration of the feature amount calculation unit 103 is the same as that of Embodiments 1 to 3, detailed description thereof will be omitted here.
- the correspondence identifying unit 114 identifies the second frame images corresponding to each of the plurality of first frame images.
- the first query moving image and the second query moving image may have different time lengths (that is, the number of first frame images and the number of second frame images may differ). Also, even when the first query video and the second query video show the same action, the timing at which the action starts and the timing at which the action ends may differ from each other.
- the correspondence identifying unit 114 identifies frame images in which a person performing a predetermined action in the first query video and a person performing a predetermined action in the second query video take the same posture. to correspond.
- the correspondence identifying unit 114 identifies the second frame images in which the person takes the same posture as the posture of the person indicated by each of the plurality of first frame images, and associates them with each other.
- a line connects the first frame image and the second frame image that correspond to each other.
- one first frame image may be associated with a plurality of second frame images.
- one second frame image may be associated with a plurality of first frame images.
- the identification of the correspondence relationship can be realized, for example, using techniques such as DTW (Dynamic Time Warping).
- DTW Dynamic Time Warping
- the distance between features Manhattan distance or Euclidean distance
- the like can be used as the distance score required to identify the correspondence.
- the integration unit 113 performs a process of integrating feature amounts calculated from each of the first frame image and the second frame image that correspond to each other.
- the integration unit 113 performs the processing on a plurality of combinations of the first frame images and the second frame images that correspond to each other, thereby creating an integrated query in which the integrated feature amounts are arranged in time series.
- FIG. 53 schematically shows an example of integrated queries. The integrated query shown is created based on the first query video and second query video shown in FIG. 51 and the correspondence relationship shown in FIG.
- the feature amounts f1 to fp corresponding to " F11 + F21” are the feature amounts f1 to fp calculated from the first frame image F11 and the feature amounts f1 to fp calculated from the second frame image F21 . It is obtained by integrating the feature quantities f 1 to f p obtained in the above.
- the feature amounts f1 to fp corresponding to " F11 + F22 " are the feature amounts f1 to fp calculated from the first frame image F11 and the feature amounts f1 to fp calculated from the second frame image F22 . It is the integration of the quantities f 1 through f p .
- the integration unit 113 can calculate an arithmetic average value or a weighted average value of the feature amounts calculated from each of the first frame image and the second frame image as the integrated feature amount.
- the feature amount f1 among the integrated feature amounts f1 to fp corresponding to " F11 + F21 " is calculated from the feature amount f1 calculated from the first frame image F11 and the second frame image F21 .
- the correspondence relationship is specified with each other, and the arithmetic mean of the feature amounts calculated from each of the plurality of frame images corresponding to each other is calculated.
- the value or weighted average value it is possible to create an integrated query in which the integrated feature values are arranged in chronological order.
- the search unit 111 searches for moving images using, as a key, an integrated query in which the integrated feature amounts f1 to fp are arranged in time series.
- an integrated query in which the integrated feature amounts f1 to fp are arranged in time series.
- the search unit 111 may adopt the techniques described in the first to third embodiments, or may adopt other techniques.
- the image processing device 100 acquires a first query moving image having a plurality of chronological first frame images and at least one second query moving image having a plurality of chronological second frame images (S500).
- the image processing apparatus 100 detects keypoints of objects included in each of the first frame image and the second frame image (S501). Next, the image processing apparatus 100 calculates feature amounts of the detected keypoints (S502).
- the image processing apparatus 100 identifies correspondence relationships between the plurality of first frame images and the plurality of second frame images (S503).
- the image processing apparatus 100 performs the process of integrating the feature amounts calculated from the first frame images and the second frame images that correspond to each other for a plurality of combinations of the first frame images and the second frame images that correspond to each other.
- an integrated query in which the integrated feature amounts are arranged in time series is created (S504).
- the illustrated processing of S600 to S604 is the same as the processing of S500 to S504 described using FIG.
- the image processing apparatus 100 searches for moving images using the integrated query created in S604 as a key (S605).
- the image processing apparatus 100 of the present embodiment which integrates a plurality of queries indicating the same posture and the same action, creates a query after integration, and performs a search using the query after integration, the posture indicated by the query This improves the search accuracy of processing for searching for images showing the same posture or action as a person or action.
- the image processing apparatus 100 of the present embodiment can be configured to employ one or more of Modifications 8 and 9 below.
- the search unit 111 may search for moving images using weighting values set for each of a plurality of types of feature amounts.
- a weighting value is set for each of the feature amounts f1 to fp , and is used to search for moving images.
- the integration unit 113 can set a weighting value for each feature amount based on the similarity of each of the plurality of types of feature amounts calculated from each of the first frame image and the second frame image that correspond to each other.
- a feature value with a low degree of similarity is a feature value related to a body part that moves characteristically in the action, and it is considered that the degree of similarity is low due to physique differences, individual differences, and the like. For this reason, the integrating unit 113 may give a higher weighting value to a feature quantity with a lower degree of similarity.
- a feature value with a high degree of similarity can also be considered as a common portion that can be the same for all people regardless of differences in physique or individual differences in behavior. For this reason, the integration unit 113 may give a higher weighting value to feature quantities with higher similarities. Which way of thinking is to be adopted can be determined based on the content of actions, required performance of search, and the like.
- the search unit 111 calculates the degree of similarity between the query image and the image to be searched for each of the plurality of types of feature amounts, integrates (for example, adds together) the plurality of degrees of similarity calculated for each of the plurality of types of feature amounts, Calculate integrated similarity.
- a weighting value can be used during this integration.
- the integrated similarity may be calculated by multiplying each of a plurality of degrees of similarity calculated for each of the plurality of types of feature amounts by a coefficient corresponding to the weighting value of each feature amount.
- the search unit 111 can search for images whose integrated similarity is greater than or equal to a threshold as images similar to the query image.
- the search unit 111 can search for videos using a part of the integrated query as a key.
- the part used as the key is the first frame included in the time period when the change in the feature value equal to or greater than the threshold (large change in the posture of the person in the video) occurs in both the first query video and the second query video. This is the part created based on the image and the second frame image.
- the time zone in which the change in the feature quantity is greater than or equal to the threshold is the time zone in which the change in the feature quantity from the predefined reference posture is greater than or equal to the threshold.
- an upright posture is defined as the reference posture
- the time period when the change from the upright posture is above the threshold is the time period when the change in the feature value is above the threshold.
- identified as The reference pose may be user definable.
- the average of postures shown in a plurality of images or the most frequent posture may be calculated as the reference posture.
- the plurality of images may be images stored in the database, images extracted from the query moving image, or images specified by the user.
- the image processing apparatus 100 detects key points of a person's body, and searches for videos showing the same actions and postures as the person's actions and postures.
- the image processing apparatus 100 may detect a keypoint of an object other than a person, and search for a video showing the same behavior and posture as the person's behavior and posture.
- Objects are not particularly limited, and examples thereof include animals, plants, natural products, artificial objects, and the like.
- the image processing apparatus 100 creates a combined query based on a plurality of query videos, and searches for videos using the created combined query.
- the image processing apparatus 100 creates a combined query based on a plurality of query still images, and searches for still images using the created combined query. Then, the image processing apparatus 100 performs the same search as the search using the weighting values set for each of the plurality of types of feature amounts described as the eighth modification in the fourth embodiment.
- FIG. 56 shows an example of a functional block diagram of the image processing apparatus 100 of this embodiment.
- the image processing apparatus 100 has a query acquisition unit 109 , a skeleton structure detection unit 102 , a feature amount calculation unit 103 , an integration unit 113 and a search unit 111 .
- the image processing apparatus 100 may further include other functional units described in the first to fourth embodiments.
- the query acquisition unit 109 acquires a first query still image and at least one second query still image. For example, the query acquisition unit 109 acquires a first query still image and a second query still image input/designated/selected by user operation.
- the first query still image and the second query still image show the same posture.
- the first query still image and the second query still image may include the same person or may include different persons. Note that it is desirable that the angle of view of the first query still image and the angle of view of the second query still image are aligned to some extent depending on the feature amount used.
- the skeletal structure detection unit 102 detects key points of the person (object) included in each of the first query still image and the second query still image. Since the configuration of the skeletal structure detection unit 102 is the same as in the first embodiment to the first embodiment, detailed description thereof is omitted here.
- the feature quantity calculation unit 103 calculates the feature quantity of the keypoints detected from each of the first query still image and the second query still image, that is, the feature quantity of the detected two-dimensional skeleton structure. Since the configuration of the feature amount calculation unit 103 is the same as that of Embodiments 1 to 4, detailed description thereof will be omitted here.
- the integration unit 113 integrates feature amounts calculated from the first query still image and the second query still image to create an integrated query.
- FIG. 57 shows a plurality of types of feature amounts f 1 to f p calculated from the first query still image, a plurality of types of feature amounts f 1 to f p calculated from the second query still image, and their integration
- FIG. 11 shows an integrated query having multiple types of feature quantities f 1 to f p that have been created.
- the integration unit 113 can calculate an arithmetic average value or a weighted average value of the feature amounts calculated from each of the first query still image and the second query still image as the integrated feature amount.
- the feature amount f1 among the feature amounts f1 to fp included in the illustrated integrated query is the feature amount f1 calculated from the first query still image and the feature amount f1 calculated from the second query still image. is the arithmetic mean or weighted mean of
- an arithmetic average value or a weighted average value of feature amounts calculated from each of the plurality of query still images is calculated. to create a federated query.
- the search unit 111 searches for still images using the integrated query as a key.
- the search unit 111 searches for still images using weighting values set for each of a plurality of types of feature amounts.
- a weighting value is set for each of the feature amounts f1 to fp included in the integrated query, and is used to search for still images.
- a method for setting weighting values and an example of search processing using weighting values are the same as those described in the fourth embodiment.
- the image processing device 100 acquires a first query still image and at least one second query still image (S700).
- the image processing apparatus 100 detects key points of objects included in each of the first query still image and the second query still image (S701). Next, the image processing apparatus 100 calculates feature amounts of the detected keypoints (S702).
- the image processing apparatus 100 integrates the feature amounts calculated from each of the first query still image and the second query still image to create an integrated query (S703). Also, the image processing apparatus 100 sets a weighting value for each of a plurality of types of feature amounts. Next, the image processing apparatus 100 searches for a still image using the integrated query created in S703 as a key and also using the multiple types of weighted values for each feature quantity created in S703 (S705).
- the image processing apparatus 100 of the present embodiment which integrates a plurality of queries indicating the same posture and performs a search using the integrated queries, the process of searching for an image showing the same posture as that indicated by the query can be performed. Improves search accuracy. In addition, it is expected that retrieval accuracy will be further improved by performing retrieval using multiple types of weighting values for each feature quantity.
- query acquiring means for acquiring a first query moving image having a plurality of time-sequential first frame images and at least one second query moving image having a plurality of time-sequential second frame images; feature quantity calculation means for calculating a feature quantity of a key point detected from an object included in each of the first frame image and the second frame image; correspondence identifying means for identifying the second frame images corresponding to each of the plurality of first frame images; A process of integrating the feature amounts calculated from the first frame images and the second frame images corresponding to each other is performed on a plurality of combinations of the first frame images and the second frame images corresponding to each other.
- an integration means for creating an integrated query in which the integrated feature values are arranged in chronological order;
- An image processing device having 2.
- 3. The image processing apparatus according to 1 or 2, further comprising search means for searching for moving images using the integrated query as a key. 4.
- the searching means searches for the moving image using a weighting value set for each of the plurality of types of feature amounts. 5. 5.
- Device. 6 The search means is The first frame image and the first frame image that are part of the integrated query and are included in a time zone in which the change in the feature amount equal to or greater than a threshold occurs in common in both the first query moving image and the second query moving image. 6.
- the image processing device according to any one of 3 to 5, wherein the moving image is searched using the part created based on the two-frame image as a key. 7.
- the computer a query acquisition step of acquiring a first query moving image having a plurality of time-series first frame images and at least one second query moving image having a plurality of time-series second frame images; a feature quantity calculation step of calculating a feature quantity of a key point detected from an object included in each of the first frame image and the second frame image; a correspondence identifying step of identifying the second frame image corresponding to each of the plurality of first frame images; A process of integrating the feature amounts calculated from the first frame images and the second frame images corresponding to each other is performed on a plurality of combinations of the first frame images and the second frame images corresponding to each other.
- An image processing method that performs 8. the computer, query acquiring means for acquiring a first query moving image having a plurality of time-sequential first frame images and at least one second query moving image having a plurality of time-sequential second frame images; feature quantity calculation means for calculating a feature quantity of a key point detected from an object included in each of the first frame image and the second frame image; correspondence identifying means for identifying the second frame images corresponding to each of the plurality of first frame images; and A process of integrating the feature amounts calculated from the first frame images and the second frame images corresponding to each other is performed on a plurality of combinations of the first frame images and the second frame images corresponding to each other.
- Integrating means for creating an integrated query in which the integrated feature values are arranged in chronological order A program that acts as a 9. query acquisition means for acquiring a first query still image and at least one second query still image; a feature amount calculation means for calculating a feature amount of a key point detected from an object included in the first query still image and the second query still image; integration means for integrating the feature amounts calculated from the first query still image and the second query still image to create an integrated query; a search means for searching still images using the integrated query as a key; has The search means is set based on the degree of similarity between each of the plurality of types of feature amounts calculated from the first query still image and each of the plurality of types of feature amounts calculated from the second query still image.
- An image processing device that searches for a still image using the weighting value for each feature quantity.
- the computer a query acquisition step of acquiring a first query still image and at least one second query still image; a feature quantity calculation step of calculating a feature quantity of a key point detected from an object included in the first query still image and the second query still image; an integration step of integrating the feature amounts calculated from the first query still image and the second query still image to create an integrated query; a search step of searching for still images using the integrated query as a key; and run In the searching step, the similarity between each of the plurality of types of feature amounts calculated from the first query still image and each of the plurality of types of feature amounts calculated from the second query still image is set based on An image processing method for retrieving a still image using the weighting value for each feature amount.
- query acquisition means for acquiring a first query still image and at least one second query still image
- feature quantity calculation means for calculating a feature quantity of a key point detected from an object included in the first query still image and the second query still image
- integration means for integrating the feature amounts calculated from the first query still image and the second query still image to create an integrated query
- Search means for searching still images using the integrated query as a key; function as The search means is set based on the degree of similarity between each of the plurality of types of feature amounts calculated from the first query still image and each of the plurality of types of feature amounts calculated from the second query still image.
- a program for retrieving a still image using the weighting value for each feature amount.
- image processing system 10 image processing device 11 skeleton detection unit 12 feature quantity calculation unit 13 recognition unit 100 image processing device 101 image acquisition unit 102 skeleton structure detection unit 103 feature quantity calculation unit 104 classification unit 105 search unit 106 input unit 107 display unit 108 height calculation unit 109 query acquisition unit 110 change calculation unit 111 search unit 112 query frame selection unit 113 integration unit 114 correspondence identification unit 200 camera 201 database 300, 301 human body model 401 two-dimensional skeletal structure
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Databases & Information Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Library & Information Science (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
Description
時系列な複数の第1フレーム画像を有する第1クエリ動画と、時系列な複数の第2フレーム画像を有する少なくとも1つの第2クエリ動画とを取得する取得するクエリ取得手段と、
前記第1フレーム画像及び前記第2フレーム画像各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段と、
複数の前記第1フレーム画像各々に対応する前記第2フレーム画像を特定する対応関係特定手段と、
互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量を統合する処理を、互いに対応する前記第1フレーム画像及び前記第2フレーム画像の複数の組み合わせに対して行うことで、統合した前記特徴量が時系列に並んだ統合クエリを作成する統合手段と、
を有する画像処理装置が提供される。
コンピュータが、
時系列な複数の第1フレーム画像を有する第1クエリ動画と、時系列な複数の第2フレーム画像を有する少なくとも1つの第2クエリ動画とを取得する取得するクエリ取得工程と、
前記第1フレーム画像及び前記第2フレーム画像各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出工程と、
複数の前記第1フレーム画像各々に対応する前記第2フレーム画像を特定する対応関係特定工程と、
互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量を統合する処理を、互いに対応する前記第1フレーム画像及び前記第2フレーム画像の複数の組み合わせに対して行うことで、統合した前記特徴量が時系列に並んだ統合クエリを作成する統合工程と、
を実行する画像処理方法が提供される。
コンピュータを、
時系列な複数の第1フレーム画像を有する第1クエリ動画と、時系列な複数の第2フレーム画像を有する少なくとも1つの第2クエリ動画とを取得する取得するクエリ取得手段、
前記第1フレーム画像及び前記第2フレーム画像各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段、
複数の前記第1フレーム画像各々に対応する前記第2フレーム画像を特定する対応関係特定手段、及び、
互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量を統合する処理を、互いに対応する前記第1フレーム画像及び前記第2フレーム画像の複数の組み合わせに対して行うことで、統合した前記特徴量が時系列に並んだ統合クエリを作成する統合手段、
として機能させるプログラムが提供される。
第1クエリ静止画と、少なくとも1つの第2クエリ静止画とを取得するクエリ取得手段と、
前記第1クエリ静止画及び前記第2クエリ静止画各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段と、
前記第1クエリ静止画及び前記第2クエリ静止画各々から算出された前記特徴量を統合し、統合クエリを作成する統合手段と、
前記統合クエリをキーとして用いて静止画を検索する検索手段と、
を有し、
前記検索手段は、前記第1クエリ静止画から算出された複数種類の前記特徴量各々と、前記第2クエリ静止画から算出された複数種類の前記特徴量各々との類似度に基づき設定された前記特徴量別の重み付け値を用いて、静止画を検索する画像処理装置が提供される。
コンピュータが、
第1クエリ静止画と、少なくとも1つの第2クエリ静止画とを取得するクエリ取得工程と、
前記第1クエリ静止画及び前記第2クエリ静止画各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出工程と、
前記第1クエリ静止画及び前記第2クエリ静止画各々から算出された前記特徴量を統合し、統合クエリを作成する統合工程と、
前記統合クエリをキーとして用いて静止画を検索する検索工程と、
を実行し、
前記検索工程では、前記第1クエリ静止画から算出された複数種類の前記特徴量各々と、前記第2クエリ静止画から算出された複数種類の前記特徴量各々との類似度に基づき設定された前記特徴量別の重み付け値を用いて、静止画を検索する画像処理方法が提供される。
コンピュータを、
第1クエリ静止画と、少なくとも1つの第2クエリ静止画とを取得するクエリ取得手段、
前記第1クエリ静止画及び前記第2クエリ静止画各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段、
前記第1クエリ静止画及び前記第2クエリ静止画各々から算出された前記特徴量を統合し、統合クエリを作成する統合手段、及び、
前記統合クエリをキーとして用いて静止画を検索する検索手段、
として機能させ、
前記検索手段は、前記第1クエリ静止画から算出された複数種類の前記特徴量各々と、前記第2クエリ静止画から算出された複数種類の前記特徴量各々との類似度に基づき設定された前記特徴量別の重み付け値を用いて、静止画を検索するプログラムが提供される。
近年、ディープラーニング等の機械学習を活用した画像認識技術が様々なシステムに応用されている。例えば、監視カメラの画像により監視を行う監視システムへの適用が進められている。監視システムに機械学習を活用することで、画像から人物の姿勢や行動等の状態をある程度把握することが可能とされつつある。
図1は、実施の形態に係る画像処理装置10の概要を示している。図1に示すように、画像処理装置10は、骨格検出部11、特徴量算出部12、及び認識部13を備えている。骨格検出部11は、カメラ等から取得される2次元画像に基づいて、複数の人物の2次元骨格構造を検出する。特徴量算出部12は、骨格検出部11により検出された複数の2次元骨格構造の特徴量を算出する。認識部13は、特徴量算出部12により算出された複数の特徴量の類似度に基づいて、複数の人物の状態の認識処理を行う。認識処理は、人物の状態の分類処理や検索処理等である。
以下、図面を参照して実施の形態1について説明する。図2は、本実施の形態に係る画像処理装置100の構成を示している。画像処理装置100は、カメラ200及びデータベース(DB)201とともに画像処理システム1を構成する。画像処理装置100を含む画像処理システム1は、画像から推定される人物の骨格構造に基づき、人物の姿勢や行動等の状態を分類及び検索するシステムである。
複数の階層による分類全身の骨格構造による分類や、上半身や下半身の骨格構造による分類、腕や脚の骨格構造による分類等を階層的に組み合わせて分類する。すなわち、骨格構造の第1の部分や第2の部分の特徴量に基づいて分類し、さらに、第1の部分や第2の部分の特徴量に重みづけを行って分類してもよい。
時系列に沿った複数枚の画像による分類時系列に連続する複数の画像における骨格構造の特徴量に基づいて分類する。例えば、時系列方向に特徴量を積み重ねて、累積値に基づいて分類してもよい。さらに、連続する複数の画像における骨格構造の特徴量の変化(変化量)に基づいて分類してもよい。
骨格構造の左右を無視した分類人物の右側と左側が反対の骨格構造を同じ骨格構造として分類する。
高さ方向の特徴量のみによる検索人物の高さ方向の特徴量のみを用いて検索することで、人物の横方向の変化の影響を抑えることができ、人物の向きや人物の体型の変化に対しロバスト性が向上する。例えば、図13の骨格構造501~503のように、人物の向きや体型が異なる場合でも、高さ方向の特徴量は大きく変化しない。このため、骨格構造501~503では、検索時(分類時)に同じ姿勢であると判断することができる。
部分検索画像において人物の体の一部が隠れている場合、認識可能な部分の情報のみを用いて検索する。例えば、図14の骨格構造511及び512のように、左足が隠れていることにより、左足のキーポイントが検出できない場合でも、検出されている他のキーポイントの特徴量を使用して検索できる。このため、骨格構造511及び512では、検索時(分類時)に同じ姿勢であると判断することができる。つまり、全てのキーポイントではなく、一部のキーポイントの特徴量を用いて、分類や検索を行うことができる。図15の骨格構造521及び522の例では、両足の向きが異なっているものの、上半身のキーポイント(A1、A2、A31、A32、A41、A42、A51、A52)の特徴量を検索クエリとすることで、同じ姿勢であると判断することができる。また、検索したい部分(特徴点)に対して、重みを付けて検索してもよいし、類似度判定の閾値を変化させてもよい。体の一部が隠れている場合、隠れた部分を無視して検索してもよいし、隠れた部分を加味して検索してもよい。隠れた部分も含めて検索することで、同じ部位が隠れているような姿勢を検索することができる。
骨格構造の左右を無視した検索人物の右側と左側が反対の骨格構造を同じ骨格構造として検索する。例えば、図16の骨格構造531及び532のように、右手を挙げている姿勢と、左手を挙げている姿勢を同じ姿勢として検索(分類)できる。図16の例では、骨格構造531と骨格構造532は、右手のキーポイントA51、右肘のキーポイントA41、左手のキーポイントA52、左肘のキーポイントA42の位置が異なるものの、その他のキーポイントの位置は同じである。骨格構造531の右手のキーポイントA51及び右肘のキーポイントA41と骨格構造532の左手のキーポイントA52及び左肘のキーポイントA42のうち、一方の骨格構造のキーポイントを左右反転させると、他方の骨格構造のキーポイントと同じ位置となり、また、骨格構造531の左手のキーポイントA52及び左肘のキーポイントA42と骨格構造532の右手のキーポイントA51及び右肘のキーポイントA41のうち、一方の骨格構造のキーポイントを左右反転させると、他方の骨格構造のキーポイントと同じ位置となるため、同じ姿勢と判断する。
縦方向と横方向の特徴量による検索人物の縦方向(Y軸方向)の特徴量のみで検索を行った後、得られた結果をさらに人物の横方向(X軸方向)の特徴量を用いて検索する。
時系列に沿った複数枚の画像による検索時系列に連続する複数の画像における骨格構造の特徴量に基づいて検索する。例えば、時系列方向に特徴量を積み重ねて、累積値に基づいて検索してもよい。さらに、連続する複数の画像における骨格構造の特徴量の変化(変化量)に基づいて検索してもよい。
以下、図面を参照して実施の形態2について説明する。本実施の形態では、実施の形態1における特徴量算出の具体例について説明する。本実施の形態では、人物の身長を用いて正規化することで特徴量を求める。その他については、実施の形態1と同様である。
具体例1では、頭部から足部までのボーンの長さを用いて身長画素数を求める。具体例1では、図20に示すように、身長算出部108は、各ボーンの長さを取得し(S211)、取得した各ボーンの長さを合計する(S212)。
具体例2では、2次元骨格構造に含まれる骨の長さと2次元画像空間上の人物の全身の長さとの関係を示す2次元骨格モデルを用いて身長画素数を求める。
具体例3では、2次元骨格構造を3次元人体モデル(3次元骨格モデル)にフィッティングさせて、フィッティングした3次元人体モデルの身長画素数を用いて全身の骨格ベクトルを求める。
図19に示すように、画像処理装置100は、身長画素数算出処理に続いて、正規化処理(S202)を行う。正規化処理では、図23に示すように、特徴量算出部103は、キーポイント高さを算出する(S241)。特徴量算出部103は、検出された骨格構造に含まれる全てのキーポイントのキーポイント高さ(画素数)を算出する。キーポイント高さは、骨格構造の最下端(例えばいずれかの足のキーポイント)からそのキーポイントまでの高さ方向の長さ(画素数)である。ここでは、一例として、キーポイント高さを、画像におけるキーポイントのY座標から求める。なお、上記のように、キーポイント高さは、カメラパラメータに基づいた鉛直投影軸に沿った方向の長さから求めてもよい。例えば、図24の例で、首のキーポイントA2の高さ(yi)は、キーポイントA2のY座標から右足のキーポイントA81または左足のキーポイントA82のY座標を引いた値である。
以下、図面を参照して実施の形態3について説明する。本実施の形態では、所望のシーンを含む動画を検索する処理の具体例について説明する。
選択処理1では、クエリフレーム選択部112は、ユーザ入力に基づきクエリフレームを選択する。すなわち、ユーザが、複数の第1フレーム画像の中の少なくとも一部をクエリフレームとして指定する入力を行う。そして、クエリフレーム選択部112は、ユーザにより指定された第1フレーム画像をクエリフレームとして選択する。
選択処理2では、クエリフレーム選択部112は、予め定められた規則に従ってクエリフレームを選択する。
選択処理3では、クエリフレーム選択部112は、予め定められた規則に従ってクエリフレームを選択する。
特徴量の変化の方向の時系列データをキーとして用いる場合、検索部111は、当該時系列データの類似度が基準値以上のDB動画を検索することができる。時系列データの類似度の算出方法は特段制限されず、あらゆる技術を採用することができる。
2つのクエリフレーム間で起きた特徴量の変化の方向をキーとして用いる場合、検索部111は、その特徴量の変化の方向を示すDB動画を検索することができる。
図45の機能ブロック図に示すように、画像処理装置100は、クエリフレーム選択部112を有さなくてもよい。この場合、変化算出部110は、隣接する第1フレーム画像間で特徴量の変化の方向を算出することができる。そして、3つ以上の第1フレーム画像を処理対象とした場合、変化算出部110は、特徴量の変化の方向の時系列な変化を示す時系列データを算出することができる。2つの第1フレーム画像のみを処理対象とした場合、変化算出部110は、その2つの画像間で起きた特徴量の変化の方向を算出することができる。
上記実施の形態では、画像処理装置100は、人物の身体のキーポイントを検出し、その変化の方向をキーとしてDB動画を検索した。変形例2では、画像処理装置100は、人物以外の物体のキーポイントを検出し、その変化の方向をキーとしてDB動画を検索することができる。物体は特段制限されず、例えば動物、植物、天然物、人工物等が例示される。
変化算出部110は、特徴量の変化の方向に加えて、特徴量の変化の大きさを算出することができる。変化算出部110は、隣接するクエリフレーム間又は隣接する第1フレーム画像間で、特徴量の変化の大きさを算出することができる。特徴量の変化の大きさは、例えば特徴量を示す数値の差分の絶対値で表すことができる。その他、特徴量の変化の大きさは、当該絶対値を規格化した値であってもよい。
変化算出部110は、特徴量の変化の方向に加えて、特徴量の変化のスピードを算出することができる。当該変形例は、図42に示すように第1フレーム画像からバラバラな間隔でクエリフレームを選択し、隣接するクエリフレーム間で特徴量の変化の方向を算出する場合に有効である。この場合、隣接するクエリフレーム間の特徴量の変化のスピードを参照することで、より類似するDB動画を検索することが可能となる。
ここまでは、検索部111は、キーとマッチングするDB動画を検索したが、キーとマッチングしないDB動画を検索してもよい。すなわち、検索部111は、キーである上記時系列データとの類似度が基準値未満のDB動画を検索してもよい。また、検索部111は、キーである特徴量の変化の方向(大きさ、スピード等を含んでもよい)を含まないDB動画を検索してもよい。
検索部111は、変化算出部110により算出された結果(特量量の変化の方向、大きさ、スピード等)に加えて、クエリ動画の第1フレーム画像の中から選択された代表画像をさらにキーとして用いてDB動画を検索することができる。代表画像は1つでもよいし、複数でもよい。例えばクエリフレームを代表画像としてもよいし、クエリフレームの中から任意の手段で選択したフレームを代表画像としてもよいし、その他の手段で第1フレーム画像の中から代表画像を選択してもよい。
・代表画像が複数である場合、より多くの代表画像と類似する(類似度が基準値以上)フレーム画像を含むDB動画の類似度を高くする。
・代表画像が複数である場合、複数の代表画像の時系列順と、複数の代表画像各々に類似するフレーム画像の時系列順とが類似するほどDB動画の類似度を高くする。
実施の形態1及び2と同様、画像処理装置100は、カメラ200及びデータベース201とともに画像処理システム1を構成してもよい。
以下、図面を参照して実施の形態4について説明する。本実施の形態では、同じ姿勢及び同じ行動を示す複数のクエリを統合し、統合後のクエリを用いて検索を行うことで、クエリが示す姿勢や行動と同じ姿勢や行動を示す画像を精度よく検索できるようにする。
検索部111は、複数種類の特徴量別に設定された重み付け値を用いて、動画を検索してもよい。図53に示す例の場合、特徴量f1乃至fp各々に重み付け値が設定され、それを用いて動画を検索することとなる。
検索部111は、統合クエリの一部分をキーとして用いて動画を検索することができる。キーとして用いる一部分は、第1クエリ動画及び第2クエリ動画両方に共通して閾値以上の特徴量の変化(動画内の人物の姿勢の大きな変化)が起きている時間帯に含まれる第1フレーム画像及び第2フレーム画像に基づき作成された部分である。閾値以上の特徴量の変化が起きている時間帯は、予め定義された基準姿勢からの特徴量の変化が閾値以上である時間帯である。例えば、基準姿勢として直立姿勢を定義すれば、直立姿勢からの変化が閾値上である時間帯、例えば手が上がったり、しゃがんだりした時間帯が、閾値以上の特徴量の変化が起きている時間帯として特定される。基準姿勢はユーザが定義できてもよい。その他、複数の画像で示される姿勢の平均や最頻姿勢が基準姿勢として算出されてもよい。複数の画像は、データベース内に記憶されている画像であってもよいし、クエリ動画から抽出された画像であってもよいし、ユーザが指定した画像であってもよい。
上記実施の形態では、画像処理装置100は、人物の身体のキーポイントを検出し、その人物の行動及び姿勢と同じ行動及び姿勢を示す動画を検索した。変形例2では、画像処理装置100は、人物以外の物体のキーポイントを検出し、その人物の行動及び姿勢と同じ行動及び姿勢を示す動画を検索してもよい。物体は特段制限されず、例えば動物、植物、天然物、人工物等が例示される。
実施の形態4では、画像処理装置100は、複数のクエリ動画に基づき結合クエリを作成し、作成した結合クエリを用いて動画を検索した。本実施の形態では、画像処理装置100は、複数のクエリ静止画に基づき結合クエリを作成し、作成した結合クエリを用いて静止画を検索する。そして、画像処理装置100は、実施の形態4で変形例8として説明した複数種類の特徴量別に設定された重み付け値を用いた検索と同様の検索を行う。
1. 時系列な複数の第1フレーム画像を有する第1クエリ動画と、時系列な複数の第2フレーム画像を有する少なくとも1つの第2クエリ動画とを取得するクエリ取得手段と、
前記第1フレーム画像及び前記第2フレーム画像各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段と、
複数の前記第1フレーム画像各々に対応する前記第2フレーム画像を特定する対応関係特定手段と、
互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量を統合する処理を、互いに対応する前記第1フレーム画像及び前記第2フレーム画像の複数の組み合わせに対して行うことで、統合した前記特徴量が時系列に並んだ統合クエリを作成する統合手段と、
を有する画像処理装置。
2. 前記統合手段は、統合した前記特徴量として、前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量の算術平均値又は加重平均値を算出する1に記載の画像処理装置。
3. 前記統合クエリをキーとして用いて動画を検索する検索手段をさらに有する1又は2に記載の画像処理装置。
4. 前記検索手段は、複数種類の前記特徴量別に設定された重み付け値を用いて、前記動画を検索する3に記載の画像処理装置。
5. 互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された複数種類の前記特徴量各々の類似度に基づき、前記特徴量別の前記重み付け値が設定される4に記載の画像処理装置。
6. 前記検索手段は、
前記統合クエリの一部分であって、前記第1クエリ動画及び前記第2クエリ動画両方に共通して閾値以上の前記特徴量の変化が起きている時間帯に含まれる前記第1フレーム画像及び前記第2フレーム画像に基づき作成された前記一部分をキーとして用いて動画を検索する3から5のいずれかに記載の画像処理装置。
7. コンピュータが、
時系列な複数の第1フレーム画像を有する第1クエリ動画と、時系列な複数の第2フレーム画像を有する少なくとも1つの第2クエリ動画とを取得するクエリ取得工程と、
前記第1フレーム画像及び前記第2フレーム画像各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出工程と、
複数の前記第1フレーム画像各々に対応する前記第2フレーム画像を特定する対応関係特定工程と、
互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量を統合する処理を、互いに対応する前記第1フレーム画像及び前記第2フレーム画像の複数の組み合わせに対して行うことで、統合した前記特徴量が時系列に並んだ統合クエリを作成する統合工程と、
を実行する画像処理方法。
8. コンピュータを、
時系列な複数の第1フレーム画像を有する第1クエリ動画と、時系列な複数の第2フレーム画像を有する少なくとも1つの第2クエリ動画とを取得するクエリ取得手段、
前記第1フレーム画像及び前記第2フレーム画像各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段、
複数の前記第1フレーム画像各々に対応する前記第2フレーム画像を特定する対応関係特定手段、及び、
互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量を統合する処理を、互いに対応する前記第1フレーム画像及び前記第2フレーム画像の複数の組み合わせに対して行うことで、統合した前記特徴量が時系列に並んだ統合クエリを作成する統合手段、
として機能させるプログラム。
9. 第1クエリ静止画と、少なくとも1つの第2クエリ静止画とを取得するクエリ取得手段と、
前記第1クエリ静止画及び前記第2クエリ静止画に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段と、
前記第1クエリ静止画及び前記第2クエリ静止画各々から算出された前記特徴量を統合し、統合クエリを作成する統合手段と、
前記統合クエリをキーとして用いて静止画を検索する検索手段と、
を有し、
前記検索手段は、前記第1クエリ静止画から算出された複数種類の前記特徴量各々と、前記第2クエリ静止画から算出された複数種類の前記特徴量各々との類似度に基づき設定された前記特徴量別の重み付け値を用いて、静止画を検索する画像処理装置。
10. コンピュータが、
第1クエリ静止画と、少なくとも1つの第2クエリ静止画とを取得するクエリ取得工程と、
前記第1クエリ静止画及び前記第2クエリ静止画に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出工程と、
前記第1クエリ静止画及び前記第2クエリ静止画各々から算出された前記特徴量を統合し、統合クエリを作成する統合工程と、
前記統合クエリをキーとして用いて静止画を検索する検索工程と、
を実行し、
前記検索工程では、前記第1クエリ静止画から算出された複数種類の前記特徴量各々と、前記第2クエリ静止画から算出された複数種類の前記特徴量各々との類似度に基づき設定された前記特徴量別の重み付け値を用いて、静止画を検索する画像処理方法。
11. コンピュータを、
第1クエリ静止画と、少なくとも1つの第2クエリ静止画とを取得するクエリ取得手段、
前記第1クエリ静止画及び前記第2クエリ静止画に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段、
前記第1クエリ静止画及び前記第2クエリ静止画各々から算出された前記特徴量を統合し、統合クエリを作成する統合手段、及び、
前記統合クエリをキーとして用いて静止画を検索する検索手段、
として機能させ、
前記検索手段は、前記第1クエリ静止画から算出された複数種類の前記特徴量各々と、前記第2クエリ静止画から算出された複数種類の前記特徴量各々との類似度に基づき設定された前記特徴量別の重み付け値を用いて、静止画を検索するプログラム。
10 画像処理装置
11 骨格検出部
12 特徴量算出部
13 認識部
100 画像処理装置
101 画像取得部
102 骨格構造検出部
103 特徴量算出部
104 分類部
105 検索部
106 入力部
107 表示部
108 身長算出部
109 クエリ取得部
110 変化算出部
111 検索部
112 クエリフレーム選択部
113 統合部
114 対応関係特定部
200 カメラ
201 データベース
300、301 人体モデル
401 2次元骨格構造
Claims (11)
- 時系列な複数の第1フレーム画像を有する第1クエリ動画と、時系列な複数の第2フレーム画像を有する少なくとも1つの第2クエリ動画とを取得するクエリ取得手段と、
前記第1フレーム画像及び前記第2フレーム画像各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段と、
複数の前記第1フレーム画像各々に対応する前記第2フレーム画像を特定する対応関係特定手段と、
互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量を統合する処理を、互いに対応する前記第1フレーム画像及び前記第2フレーム画像の複数の組み合わせに対して行うことで、統合した前記特徴量が時系列に並んだ統合クエリを作成する統合手段と、
を有する画像処理装置。 - 前記統合手段は、統合した前記特徴量として、前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量の算術平均値又は加重平均値を算出する請求項1に記載の画像処理装置。
- 前記統合クエリをキーとして用いて動画を検索する検索手段をさらに有する請求項1又は2に記載の画像処理装置。
- 前記検索手段は、複数種類の前記特徴量別に設定された重み付け値を用いて、前記動画を検索する請求項3に記載の画像処理装置。
- 互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された複数種類の前記特徴量各々の類似度に基づき、前記特徴量別の前記重み付け値が設定される請求項4に記載の画像処理装置。
- 前記検索手段は、
前記統合クエリの一部分であって、前記第1クエリ動画及び前記第2クエリ動画両方に共通して閾値以上の前記特徴量の変化が起きている時間帯に含まれる前記第1フレーム画像及び前記第2フレーム画像に基づき作成された前記一部分をキーとして用いて動画を検索する請求項3から5のいずれか1項に記載の画像処理装置。 - コンピュータが、
時系列な複数の第1フレーム画像を有する第1クエリ動画と、時系列な複数の第2フレーム画像を有する少なくとも1つの第2クエリ動画とを取得するクエリ取得工程と、
前記第1フレーム画像及び前記第2フレーム画像各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出工程と、
複数の前記第1フレーム画像各々に対応する前記第2フレーム画像を特定する対応関係特定工程と、
互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量を統合する処理を、互いに対応する前記第1フレーム画像及び前記第2フレーム画像の複数の組み合わせに対して行うことで、統合した前記特徴量が時系列に並んだ統合クエリを作成する統合工程と、
を実行する画像処理方法。 - コンピュータを、
時系列な複数の第1フレーム画像を有する第1クエリ動画と、時系列な複数の第2フレーム画像を有する少なくとも1つの第2クエリ動画とを取得するクエリ取得手段、
前記第1フレーム画像及び前記第2フレーム画像各々に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段、
複数の前記第1フレーム画像各々に対応する前記第2フレーム画像を特定する対応関係特定手段、及び、
互いに対応する前記第1フレーム画像及び前記第2フレーム画像各々から算出された前記特徴量を統合する処理を、互いに対応する前記第1フレーム画像及び前記第2フレーム画像の複数の組み合わせに対して行うことで、統合した前記特徴量が時系列に並んだ統合クエリを作成する統合手段、
として機能させるプログラム。 - 第1クエリ静止画と、少なくとも1つの第2クエリ静止画とを取得するクエリ取得手段と、
前記第1クエリ静止画及び前記第2クエリ静止画に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段と、
前記第1クエリ静止画及び前記第2クエリ静止画各々から算出された前記特徴量を統合し、統合クエリを作成する統合手段と、
前記統合クエリをキーとして用いて静止画を検索する検索手段と、
を有し、
前記検索手段は、前記第1クエリ静止画から算出された複数種類の前記特徴量各々と、前記第2クエリ静止画から算出された複数種類の前記特徴量各々との類似度に基づき設定された前記特徴量別の重み付け値を用いて、静止画を検索する画像処理装置。 - コンピュータが、
第1クエリ静止画と、少なくとも1つの第2クエリ静止画とを取得するクエリ取得工程と、
前記第1クエリ静止画及び前記第2クエリ静止画に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出工程と、
前記第1クエリ静止画及び前記第2クエリ静止画各々から算出された前記特徴量を統合し、統合クエリを作成する統合工程と、
前記統合クエリをキーとして用いて静止画を検索する検索工程と、
を実行し、
前記検索工程では、前記第1クエリ静止画から算出された複数種類の前記特徴量各々と、前記第2クエリ静止画から算出された複数種類の前記特徴量各々との類似度に基づき設定された前記特徴量別の重み付け値を用いて、静止画を検索する画像処理方法。 - コンピュータを、
第1クエリ静止画と、少なくとも1つの第2クエリ静止画とを取得するクエリ取得手段、
前記第1クエリ静止画及び前記第2クエリ静止画に含まれる物体から検出されたキーポイントの特徴量を算出する特徴量算出手段、
前記第1クエリ静止画及び前記第2クエリ静止画各々から算出された前記特徴量を統合し、統合クエリを作成する統合手段、及び、
前記統合クエリをキーとして用いて静止画を検索する検索手段、
として機能させ、
前記検索手段は、前記第1クエリ静止画から算出された複数種類の前記特徴量各々と、前記第2クエリ静止画から算出された複数種類の前記特徴量各々との類似度に基づき設定された前記特徴量別の重み付け値を用いて、静止画を検索するプログラム。
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US18/275,765 US20240126806A1 (en) | 2021-05-26 | 2021-05-26 | Image processing apparatus, and image processing method |
PCT/JP2021/019993 WO2022249331A1 (ja) | 2021-05-26 | 2021-05-26 | 画像処理装置、画像処理方法、およびプログラム |
JP2023523804A JPWO2022249331A1 (ja) | 2021-05-26 | 2021-05-26 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/JP2021/019993 WO2022249331A1 (ja) | 2021-05-26 | 2021-05-26 | 画像処理装置、画像処理方法、およびプログラム |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2022249331A1 true WO2022249331A1 (ja) | 2022-12-01 |
Family
ID=84228654
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/JP2021/019993 WO2022249331A1 (ja) | 2021-05-26 | 2021-05-26 | 画像処理装置、画像処理方法、およびプログラム |
Country Status (3)
Country | Link |
---|---|
US (1) | US20240126806A1 (ja) |
JP (1) | JPWO2022249331A1 (ja) |
WO (1) | WO2022249331A1 (ja) |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2020135747A (ja) * | 2019-02-25 | 2020-08-31 | 株式会社日立ソリューションズ | 行動分析装置および行動分析方法 |
CN109308438B (zh) * | 2017-07-28 | 2020-11-27 | 上海形趣信息科技有限公司 | 动作识别库的建立方法、电子设备、存储介质 |
-
2021
- 2021-05-26 US US18/275,765 patent/US20240126806A1/en active Pending
- 2021-05-26 WO PCT/JP2021/019993 patent/WO2022249331A1/ja active Application Filing
- 2021-05-26 JP JP2023523804A patent/JPWO2022249331A1/ja active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109308438B (zh) * | 2017-07-28 | 2020-11-27 | 上海形趣信息科技有限公司 | 动作识别库的建立方法、电子设备、存储介质 |
JP2020135747A (ja) * | 2019-02-25 | 2020-08-31 | 株式会社日立ソリューションズ | 行動分析装置および行動分析方法 |
Also Published As
Publication number | Publication date |
---|---|
JPWO2022249331A1 (ja) | 2022-12-01 |
US20240126806A1 (en) | 2024-04-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP7556556B2 (ja) | 画像処理装置、画像処理方法及び画像処理プログラム | |
JP7409499B2 (ja) | 画像処理装置、画像処理方法、及びプログラム | |
JP7416252B2 (ja) | 画像処理装置、画像処理方法、及びプログラム | |
JP2024103572A (ja) | 情報処理装置、情報処理方法、およびプログラム | |
JP7364077B2 (ja) | 画像処理装置、画像処理方法、及びプログラム | |
JP7435781B2 (ja) | 画像選択装置、画像選択方法、及びプログラム | |
JP7491380B2 (ja) | 画像選択装置、画像選択方法、及びプログラム | |
WO2022079794A1 (ja) | 画像選択装置、画像選択方法、及びプログラム | |
WO2022249331A1 (ja) | 画像処理装置、画像処理方法、およびプログラム | |
WO2022249278A1 (ja) | 画像処理装置、画像処理方法、およびプログラム | |
JP7396364B2 (ja) | 画像処理装置、画像処理方法及び画像処理プログラム | |
JP7302741B2 (ja) | 画像選択装置、画像選択方法、およびプログラム | |
JP7501621B2 (ja) | 画像選択装置、画像選択方法、およびプログラム | |
JP7375921B2 (ja) | 画像分類装置、画像分類方法、およびプログラム | |
JP7435754B2 (ja) | 画像選択装置、画像選択方法、及びプログラム | |
JP7485040B2 (ja) | 画像処理装置、画像処理方法、及びプログラム | |
JP7468642B2 (ja) | 画像処理装置、画像処理方法、及びプログラム | |
WO2022079795A1 (ja) | 画像選択装置、画像選択方法、及びプログラム | |
WO2023152974A1 (ja) | 画像処理装置、画像処理方法、およびプログラム | |
WO2023152977A1 (ja) | 画像処理装置、画像処理方法、およびプログラム | |
WO2023152841A1 (ja) | 画像処理システム、画像処理方法及び非一時的なコンピュータ可読媒体 | |
WO2023152971A1 (ja) | 画像処理装置、画像処理方法、およびプログラム | |
WO2023112321A1 (ja) | 画像処理システム、画像処理方法及び非一時的なコンピュータ可読媒体 | |
WO2023089690A1 (ja) | 検索装置、検索方法、およびプログラム | |
WO2023152973A1 (ja) | 画像処理装置、画像処理方法、およびプログラム |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 21942985 Country of ref document: EP Kind code of ref document: A1 |
|
WWE | Wipo information: entry into national phase |
Ref document number: 18275765 Country of ref document: US |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2023523804 Country of ref document: JP |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 21942985 Country of ref document: EP Kind code of ref document: A1 |