CN112017171B - Image processing index evaluation method, system, equipment and medium - Google Patents

Image processing index evaluation method, system, equipment and medium Download PDF

Info

Publication number
CN112017171B
CN112017171B CN202010879122.8A CN202010879122A CN112017171B CN 112017171 B CN112017171 B CN 112017171B CN 202010879122 A CN202010879122 A CN 202010879122A CN 112017171 B CN112017171 B CN 112017171B
Authority
CN
China
Prior art keywords
data
index evaluation
track
rate
image processing
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202010879122.8A
Other languages
Chinese (zh)
Other versions
CN112017171A (en
Inventor
周曦
姚志强
徐飞
刘盛中
焦宾
龚强
侯永顺
周鑫
张政
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Sichuan Yuncong Tianfu Artificial Intelligence Technology Co Ltd
Original Assignee
Sichuan Yuncong Tianfu Artificial Intelligence Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Sichuan Yuncong Tianfu Artificial Intelligence Technology Co Ltd filed Critical Sichuan Yuncong Tianfu Artificial Intelligence Technology Co Ltd
Priority to CN202010879122.8A priority Critical patent/CN112017171B/en
Publication of CN112017171A publication Critical patent/CN112017171A/en
Application granted granted Critical
Publication of CN112017171B publication Critical patent/CN112017171B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Human Resources & Organizations (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Biophysics (AREA)
  • Mathematical Physics (AREA)
  • Biomedical Technology (AREA)
  • Artificial Intelligence (AREA)
  • Quality & Reliability (AREA)
  • Strategic Management (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Development Economics (AREA)
  • Educational Administration (AREA)
  • Economics (AREA)
  • Computational Linguistics (AREA)
  • Game Theory and Decision Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Image Analysis (AREA)

Abstract

The invention provides an image processing index evaluation method, system, device and medium, comprising the following steps: acquiring image data, and outputting structured data after one or more preset algorithm model operations; analyzing the structured data and predicting track data corresponding to one or more target objects respectively according to the input marking data; evaluating the predicted trajectory data and outputting an index evaluation result; the method is beneficial to continuous optimization of the algorithm through quantifiable indexes, and the evaluation result is more accurate and credible.

Description

Image processing index evaluation method, system, equipment and medium
Technical Field
The invention relates to the field of artificial intelligence, in particular to an image processing index evaluation method, system, equipment and medium.
Background
With the progress of artificial intelligence technology into life, deep study of deep learning algorithms, support of hardware equipment, more and more scenes and industries begin to use artificial intelligence equipment, and a fully-structured system with high aggregation, high integration and low coupling begins to face the market, but now a systematic algorithm effect evaluation method and system and key quantitative indexes are still lacking.
Disclosure of Invention
In view of the problems in the prior art, the invention provides an image processing index evaluation method, system, device and medium, which mainly solve the problem that the existing algorithm lacks structured evaluation.
In order to achieve the above and other objects, the present invention adopts the following technical solutions.
An image processing index evaluation method includes:
acquiring image data, and outputting structured data after one or more preset algorithm model operations;
analyzing the structured data and predicting track data corresponding to one or more target objects respectively according to the input marking data;
and evaluating the predicted trajectory data and outputting an index evaluation result.
Optionally, according to the input labeled data, analyzing the structured data and predicting trajectory data corresponding to one or more target objects respectively: the labeling data comprises label information of the target object; and performing label determination on the structured data according to the label information and a preset label determination algorithm, and obtaining track data corresponding to the same target object based on the determined label information.
Optionally, the preset tag determination algorithm includes a cross-over method, and the tag information includes a position, an ID, and/or attribute information of the target object.
Optionally, evaluating the predicted trajectory data and outputting an index evaluation result: and evaluating the track data according to a preset evaluation algorithm of one or more indexes.
Optionally, the index evaluation result comprises at least one or more of the following: recall rate, snapshot repetition rate, snapshot accuracy rate, false capture rate, target association rate, false association rate, missed association rate, and attribute accuracy rate.
Optionally, the target association rate is calculated as follows:
Figure BDA0002653564510000021
wherein the content of the first and second substances,
Figure BDA0002653564510000022
Piwhether the ith point in the track data belongs to the track of the same target object or not is represented, the value is 0 or 1, and the ith point does not belong to or belongs to the track respectively; t represents a fault-tolerant coefficient, and the value of T is between 0 and 1; n represents the number of points in the track data, m represents the number of moving tracks, SkAnd whether the Kth track has an incidence relation with other tracks or not is shown, the value is 0 or 1, and no incidence relation or incidence relation is shown respectively.
Optionally, the error rate is calculated as follows:
Figure BDA0002653564510000023
wherein the content of the first and second substances,
Figure BDA0002653564510000024
Piwhether the ith point in the track data belongs to the track of the same target object or not is represented, the value is 0 or 1, and the ith point does not belong to or belongs to the track respectively; t represents a fault-tolerant coefficient, and the value of T is between 0 and 1; n represents the number of points in the track data, m represents the number of moving tracks, SkAnd whether the Kth track has an incidence relation with other tracks or not is shown, the value is 0 or 1, and no incidence relation or incidence relation is shown respectively.
Optionally, the leak rate is calculated as follows:
LRR＝1–RRR
wherein RRR represents a target association rate.
Optionally, the attribute accuracy is calculated as follows:
Figure BDA0002653564510000025
wherein the content of the first and second substances,
Figure BDA0002653564510000031
aiwhether the ith point in the track data belongs to the same track data or not is represented, the value is 0 or 1, Tn represents a fault-tolerant coefficient, and the value is between 0 and 1; n represents the number of points in the track data, M represents the number of the track data, and M represents the total number of the track data or the target objects.
Optionally, the recall rate is calculated as follows:
RecallR＝PRO/LO
wherein, PRO represents the number of target objects predicted to be correct after the repetition; and LO represents the number of labeling target objects.
Optionally, the repetition rate is calculated as follows:
RepateR＝(PRT-ES)/LO–1
wherein PRT represents the total number of target objects with correct prediction; ES denotes the number of false beats.
Optionally, the snapshot accuracy is calculated as follows:
SRR＝(PRT-ES)/PRT
wherein PRT represents the total number of target objects with correct prediction; ES denotes the number of false beats.
Optionally, the false capture rate is calculated as follows:
SER＝ES/PRT
wherein PRT represents the total number of target objects with correct prediction; ES denotes the number of false beats.
An image processing index evaluation system comprising:
the system comprises a structuring module, a data processing module and a data processing module, wherein the structuring module is used for acquiring image data, and outputting the structured data after one or more preset algorithm model operations;
the track acquisition module is used for analyzing the structured data and predicting track data corresponding to one or more target objects according to the input marking data;
and the index evaluation module is used for evaluating the predicted track data and outputting an index evaluation result.
Optionally, the track obtaining module includes a tag determining unit, configured to determine that the annotation data includes tag information of the target object; and performing label determination on the structured data according to the label information and a preset label determination algorithm, and obtaining track data corresponding to the same target object based on the determined label information.
Optionally, the preset tag determination algorithm includes a cross-over method, and the tag information includes a position, an ID, and/or attribute information of the target object.
Optionally, the index evaluation module includes an algorithm evaluation unit, configured to evaluate the trajectory data according to a preset one or more index evaluation algorithms.
Optionally, the index evaluation result comprises at least one or more of the following: recall rate, snapshot repetition rate, snapshot accuracy rate, false capture rate, target association rate, false association rate, missed association rate, and attribute accuracy rate.
An apparatus, comprising:
one or more processors; and
one or more machine readable media having instructions stored thereon that, when executed by the one or more processors, cause the apparatus to perform the image processing metric evaluation method.
One or more machine readable media having instructions stored thereon that, when executed by one or more processors, cause an apparatus to perform the image processing metric evaluation method.
As described above, the image processing index evaluation method, system, device, and medium of the present invention have the following advantageous effects.
And obtaining the moving track information through the marked target object and the target object position in the image sequence, further calculating a track evaluation index for quantifying the algorithm performance, and obtaining the evaluation output of the structured algorithm effect.
Drawings
Fig. 1 is a flowchart of an image processing index evaluation method according to an embodiment of the invention.
FIG. 2 is a block diagram of an image processing index evaluation system according to an embodiment of the present invention.
Fig. 3 is a schematic structural diagram of a terminal device in an embodiment of the present invention.
Fig. 4 is a schematic structural diagram of a terminal device in another embodiment of the present invention.
Detailed Description
The embodiments of the present invention are described below with reference to specific embodiments, and other advantages and effects of the present invention will be easily understood by those skilled in the art from the disclosure of the present specification. The invention is capable of other and different embodiments and of being practiced or of being carried out in various ways, and its several details are capable of modification in various respects, all without departing from the spirit and scope of the present invention. It is to be noted that the features in the following embodiments and examples may be combined with each other without conflict.
It should be noted that the drawings provided in the following embodiments are only for illustrating the basic idea of the present invention, and the components related to the present invention are only shown in the drawings rather than drawn according to the number, shape and size of the components in actual implementation, and the type, quantity and proportion of the components in actual implementation may be changed freely, and the layout of the components may be more complicated.
Referring to FIG. 1, the present invention provides an image processing index evaluation method, which includes steps S01-S03.
In step S01, image data is acquired, and structured data is output after one or more preset algorithm model operations:
in one embodiment, when image data is acquired, an image sequence meeting preset scene conditions can be acquired, one or more frames of images are selected from the image sequence for target object labeling, and labeled data is acquired; wherein the scene condition comprises at least one of: weather, lighting, target object density;
in one embodiment, the method can be selectively applied to different scenes, such as traffic roads, stations, airports and other scenes with large pedestrian or vehicle flow. Taking a traffic road scene as an example, a plurality of image acquisition devices (such as cameras) can be used for acquiring video streams at different positions of a certain road section, or acquiring video streams at different angles of the same traffic intersection, and converting video stream information into an image sequence. The video streams collected by the plurality of image devices can be data of different time periods, and the obtained image sequences meet preset scene conditions. Wherein the scene conditions may include one or more of weather, lighting, traffic or people stream density, and the like. For example, video stream data which is acquired every other day and meets a certain illumination condition at the same traffic intersection or a plurality of traffic intersections can be acquired through the camera, and then an image sequence is acquired.
In an embodiment, when the image sequence is acquired by a plurality of image acquisition devices, one frame of image can be selected from the image sequence acquired by each image acquisition device respectively for target object annotation, and annotation data is acquired. The label data may include label information of the target object, such as a category, a label location, a target object number, a target object attribute, and the like. For example, the target object may include a person, a motor vehicle, a non-motor vehicle, and the like, and the target object attribute may include information on a male, a female, a child, an old person, a vehicle type, a license plate, and the like.
In an embodiment, a deep learning algorithm (such as a deep neural network algorithm, a convolutional neural network, or the like) may be used to perform tracking detection on a target object in an image sequence, and structured data is obtained after processing such as key point alignment, target object association, image quality score evaluation, feature extraction, and attribute analysis.
Specifically, an image quality prediction model can be constructed through a BP neural network, a large number of images under the same application scene are collected to serve as neural network input, the subjective observation result score of human eyes serves as a training target, the neural network is supervised and learned, and the obtained prediction model is used for performing quality scoring on each frame of image in an image sequence. And selecting the image with the quality score reaching a set score threshold value for feature processing. The position of the target object in the image may be further obtained by using a key point alignment method. Taking the traffic flow on the road as an example, an ASM method is adopted, a continuous closed curve composed of n control points is used as a snake model, an energy function is used as a matching degree evaluation function, the model is set around the estimated position of the target vehicle, and the energy function is minimized through continuous iteration to obtain the boundary and key features of the target vehicle. Finally, tracking and detecting the target object through feature comparison in each frame of image, classifying the same target object in the image sequence into one class, and giving the same label information (such as ID and the like). To ensure that different target objects have different tag information.
For example, the structured data can be represented as:
CLi (automotive label) or NCLi (non-automotive label) or PLi (human face, human body, human head label).
In step S02, analyzing the structured data and predicting trajectory data corresponding to one or more target objects according to the input labeled data;
in an embodiment, the target objects in the image sequence may be associated according to a preset association rule, and an association tag may be set. Wherein the association rule may include: the dependency relationship between target objects, the positional relationship between target objects, the following relationship between target objects, etc. Can be adjusted according to specific application scenes. If the vehicle and the face image in the vehicle are detected simultaneously, the vehicle is determined to be associated with the face, and if the vehicle tag is ID1 and the face tag is ID3, the two associated tags can be set to ID1-ID3, and the associated tags are recorded in the vehicle identification set. For another example, whether a plurality of vehicles and/or a plurality of persons follow or are in a same-row relationship can be judged according to the positions of the vehicles or the persons in the multi-frame images, and then the associated labels are set for the vehicles and/or the persons in the same row.
In an embodiment, the cross ratio between the labeling position of the target object in the labeling data and the position of the target object near the corresponding position in the image sequence may be calculated, and the ratio of the intersection and the union of the two position areas is obtained as the cross ratio. Further, a crossing ratio threshold value can be set, and when the crossing ratio of the annotation position to the position of the target object corresponding to one or more frames of images in the image sequence is greater than the crossing ratio threshold value, the positions of the target objects meeting the crossing ratio threshold value are combined into an initial track set. The crossing ratio threshold value can be set according to actual conditions, and can be set to be greater than 50% for example. And respectively calculating the intersection ratio of the positions of other target objects in the image sequence and the position points in the initial movement track set, if the intersection ratio is more than 50%, merging the corresponding positions into the initial track set, and gradually acquiring track data corresponding to the target objects.
In an embodiment, the categories of target objects in the image sequence may be predicted by using an intersection ratio threshold, for example, according to an iou (intersection over union) intersection ratio method, target objects meeting the intersection ratio threshold are determined as the same category, so as to confirm that the most probable (maximum IoU principle) target object in the jth image sequence is CLj (motor vehicle), NCLj (non-motor vehicle), or PLj (human face, human body, human head, etc.) tag, and then the movement trajectory of the kth target object may be recorded as:
uk { CkL1, ckl2.. CkLn } or
{ NCkL1, NCkL2.. NCkLn } or
{PkL1,PkL2...PkLn}
In the image sequences acquired by the multiple image acquisition devices, multiple discontinuous track data of the same target object may exist in different time periods, and the multiple track data of the same target object can be stored in an associated manner.
In step S03, the predicted trajectory data is evaluated and an index evaluation result is output.
In one embodiment, the same target object is endowed with the same ID through tracking detection of the target object in the image sequence, and the target object ID corresponding to the position point in the ID identification track data is used for distinguishing, judging whether the position point corresponds to the same target object or not, and further calculating the index evaluation result. The index evaluation result can include the accuracy of the snapshot target object, the association rate of the target object, the error rate of the target object, the missing rate of the target object, the recall rate, the repetition rate, the attribute accuracy and the like.
Specifically, two functions with a value of 0 or 1 may be defined for calculating the trajectory estimation indicator.
Suppose that
Figure BDA0002653564510000071
Then
Target association rate:
Figure BDA0002653564510000072
the error association rate:
Figure BDA0002653564510000073
wherein: piWhether the ith point in the track data belongs to the track data corresponding to the same target object or not is represented, the value is 0 or 1, and the ith point does not belong to or belongs to the track data respectively; t represents a fault-tolerant coefficient (the value is 0-1), the larger the value is, the higher the requirement is, n represents the number of the middle points of the track data, m represents the number of the moving tracks, and S representskIt indicates whether the kth track has an association relationship with other tracks, and can only be 0 or 1.
The leakage correlation rate: LRR 1-RRR
Attribute accuracy:
Figure BDA0002653564510000074
wherein: a isiWhether the ith point in the track data belongs to the same track data is only 0 or 1, Tn represents a fault-tolerant coefficient and takes a value between 0 and 1; n represents the number of points in the trajectory data, M represents the number of trajectory data, and M represents the total number of trajectory data or target objects.
The recall ratio is as follows: RecallR ═ PRO/LO
Repetition rate: RepateR ═ PRT-ES/LO-1
The snapshot accuracy is as follows: SRR ═ PRT-ES)/PRT
The false grabbing rate is as follows: SER ═ ES/PRT
Wherein: PRO-number of correct target objects in the predicted image sequence (deduplicated); LO-marking the number of target objects; PRT-predicted correct total; ES-number of false beats.
And the index evaluation result is stored, so that subsequent checking and problem positioning are facilitated.
Specifically, the difference of the index evaluation results obtained by setting different IOU thresholds can be compared according to the set IOU threshold statistical index evaluation result, so that the overall performance of the algorithm can be quantitatively evaluated.
Referring to fig. 2, the present embodiment provides an image processing index evaluation system for performing the image processing index evaluation method in the foregoing method embodiment. Since the technical principle of the system embodiment is similar to that of the method embodiment, repeated description of the same technical details is omitted.
In an embodiment, the image processing index evaluation system includes a structuring module 10, a trajectory acquisition module 11 and an index evaluation module 12, the annotation data acquisition module 10 is configured to assist in executing the step S01 described in the foregoing method embodiment, and the trajectory acquisition module 11 is configured to execute the step S02 described in the foregoing method embodiment; the trajectory evaluation module 12 is configured to perform step S03 described in the foregoing method embodiment.
In an embodiment, the trajectory acquisition module includes a tag determination unit, configured to make the annotation data include tag information of the target object; and performing label determination on the structured data according to the label information and a preset label determination algorithm, and obtaining track data corresponding to the same target object based on the determined label information.
In one embodiment, the preset tag determination algorithm comprises a cross-comparison method, and the tag information comprises the position, the ID and/or the attribute information of the target object.
In one embodiment, the index evaluation module includes an algorithm evaluation unit for evaluating the trajectory data according to one or more preset index evaluation algorithms.
In one embodiment, the index evaluation result includes at least one or more of the following: recall rate, snapshot repetition rate, snapshot accuracy rate, false capture rate, target association rate, false association rate, missed association rate, and attribute accuracy rate.
An embodiment of the present application further provides an apparatus, which may include: one or more processors; and one or more machine readable media having instructions stored thereon that, when executed by the one or more processors, cause the apparatus to perform the method of fig. 1. In practical applications, the device may be used as a terminal device, and may also be used as a server, where examples of the terminal device may include: the mobile terminal includes a smart phone, a tablet computer, an electronic book reader, an MP3 (Moving Picture Experts Group Audio Layer III) player, an MP4 (Moving Picture Experts Group Audio Layer IV) player, a laptop, a vehicle-mounted computer, a desktop computer, a set-top box, an intelligent television, a wearable device, and the like.
The embodiment of the present application further provides a non-volatile readable storage medium, where one or more modules (programs) are stored in the storage medium, and when the one or more modules are applied to a device, the device may execute instructions (instructions) of steps included in the image processing index evaluation method in fig. 1 according to the embodiment of the present application.
Fig. 3 is a schematic diagram of a hardware structure of a terminal device according to an embodiment of the present application. As shown, the terminal device may include: an input device 1100, a first processor 1101, an output device 1102, a first memory 1103, and at least one communication bus 1104. The communication bus 1104 is used to implement communication connections between the elements. The first memory 1103 may include a high-speed RAM memory, and may also include a non-volatile storage NVM, such as at least one disk memory, and the first memory 1103 may store various programs for performing various processing functions and implementing the method steps of the present embodiment.
Alternatively, the first processor 1101 may be, for example, a Central Processing Unit (CPU), an Application Specific Integrated Circuit (ASIC), a Digital Signal Processor (DSP), a Digital Signal Processing Device (DSPD), a Programmable Logic Device (PLD), a Field Programmable Gate Array (FPGA), a controller, a microcontroller, a microprocessor, or other electronic components, and the processor 1101 is coupled to the input device 1100 and the output device 1102 through a wired or wireless connection.
Optionally, the input device 1100 may include a variety of input devices, such as at least one of a user-oriented user interface, a device-oriented device interface, a software programmable interface, a camera, and a sensor. Optionally, the device interface facing the device may be a wired interface for data transmission between devices, or may be a hardware plug-in interface (e.g., a USB interface, a serial port, etc.) for data transmission between devices; optionally, the user-facing user interface may be, for example, a user-facing control key, a voice input device for receiving voice input, and a touch sensing device (e.g., a touch screen with a touch sensing function, a touch pad, etc.) for receiving user touch input; optionally, the programmable interface of the software may be, for example, an entry for a user to edit or modify a program, such as an input pin interface or an input interface of a chip; the output devices 1102 may include output devices such as a display, audio, and the like.
In this embodiment, the processor of the terminal device includes a function for executing each module of the speech recognition apparatus in each device, and specific functions and technical effects may refer to the above embodiments, which are not described herein again.
Fig. 4 is a schematic hardware structure diagram of a terminal device according to another embodiment of the present application. Fig. 4 is a specific embodiment of fig. 3 in an implementation process. As shown, the terminal device of the present embodiment may include a second processor 1201 and a second memory 1202.
The second processor 1201 executes the computer program code stored in the second memory 1202 to implement the method described in fig. 1 in the above embodiment.
The second memory 1202 is configured to store various types of data to support operations at the terminal device. Examples of such data include instructions for any application or method operating on the terminal device, such as messages, pictures, videos, and so forth. The second memory 1202 may include a Random Access Memory (RAM) and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.
Optionally, the first processor 1201 is provided in the processing assembly 1200. The terminal device may further include: communication component 1203, power component 1204, multimedia component 1205, speech component 1206, input/output interfaces 1207, and/or sensor component 1208. The specific components included in the terminal device are set according to actual requirements, which is not limited in this embodiment.
The processing component 1200 generally controls the overall operation of the terminal device. The processing assembly 1200 may include one or more second processors 1201 to execute instructions to perform all or part of the steps of the method illustrated in fig. 1 described above. Further, the processing component 1200 can include one or more modules that facilitate interaction between the processing component 1200 and other components. For example, the processing component 1200 can include a multimedia module to facilitate interaction between the multimedia component 1205 and the processing component 1200.
The power supply component 1204 provides power to the various components of the terminal device. The power components 1204 may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for the terminal device.
The multimedia components 1205 include a display screen that provides an output interface between the terminal device and the user. In some embodiments, the display screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the display screen includes a touch panel, the display screen may be implemented as a touch screen to receive an input signal from a user. The touch panel includes one or more touch sensors to sense touch, slide, and gestures on the touch panel. The touch sensor may not only sense the boundary of a touch or slide action, but also detect the duration and pressure associated with the touch or slide operation.
The voice component 1206 is configured to output and/or input voice signals. For example, the voice component 1206 includes a Microphone (MIC) configured to receive external voice signals when the terminal device is in an operational mode, such as a voice recognition mode. The received speech signal may further be stored in the second memory 1202 or transmitted via the communication component 1203. In some embodiments, the speech component 1206 further comprises a speaker for outputting speech signals.
The input/output interface 1207 provides an interface between the processing component 1200 and peripheral interface modules, which may be click wheels, buttons, etc. These buttons may include, but are not limited to: a volume button, a start button, and a lock button.
The sensor component 1208 includes one or more sensors for providing various aspects of status assessment for the terminal device. For example, the sensor component 1208 may detect an open/closed state of the terminal device, relative positioning of the components, presence or absence of user contact with the terminal device. The sensor assembly 1208 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact, including detecting the distance between the user and the terminal device. In some embodiments, the sensor assembly 1208 may also include a camera or the like.
The communication component 1203 is configured to facilitate communications between the terminal device and other devices in a wired or wireless manner. The terminal device may access a wireless network based on a communication standard, such as WiFi, 2G or 3G, or a combination thereof. In one embodiment, the terminal device may include a SIM card slot therein for inserting a SIM card therein, so that the terminal device may log onto a GPRS network to establish communication with the server via the internet.
As can be seen from the above, the communication component 1203, the voice component 1206, the input/output interface 1207 and the sensor component 1208 referred to in the embodiment of fig. 4 can be implemented as the input device in the embodiment of fig. 3.
In summary, the image processing index evaluation method, system, device and medium of the present invention facilitate continuous optimization of the algorithm by constructing quantifiable image evaluation indexes and trajectory evaluation indexes; the moving track information is obtained by marking the intersection ratio of the target object and the target object in the snapshot image, so that the evaluation result is more accurate and credible; and the algorithm effect evaluation is automatically carried out, the automation level is improved, the labor input is less, and the efficiency is improved. Therefore, the invention effectively overcomes various defects in the prior art and has high industrial utilization value.
The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.

Claims (16)

1. An image processing index evaluation method is characterized by comprising the following steps:
acquiring image data, and outputting structured data after one or more preset algorithm model operations;
analyzing the structured data and predicting track data corresponding to one or more target objects respectively according to the input marking data; the labeling data comprises label information of the target object; performing label determination on the structured data according to the label information and a preset label determination algorithm, and obtaining track data corresponding to the same target object based on the determined label information; the preset label determination algorithm comprises a cross ratio method, and the label information comprises the position, ID and/or attribute information of a target object;
and evaluating the predicted trajectory data and outputting an index evaluation result.
2. The image processing index evaluation method according to claim 1, wherein the trajectory data is evaluated according to a preset index evaluation algorithm or algorithms;
the algorithmic model includes one or more of: detection tracking, key point alignment, target association, quality score evaluation, feature extraction and attribute analysis.
3. The image processing index evaluation method according to claim 1, wherein the index evaluation result includes at least one or more of: recall rate, snapshot repetition rate, snapshot accuracy rate, false capture rate, target association rate, false association rate, missed association rate, and attribute accuracy rate.
4. The image processing index evaluation method according to claim 3, wherein the target association ratio is calculated as follows:
Figure FDA0003236726240000011
wherein the content of the first and second substances,
Figure FDA0003236726240000012
Piwhether the ith point in the track data belongs to the track of the same target object or not is represented, the value is 0 or 1, and the ith point does not belong to or belongs to the track respectively; t represents a fault-tolerant coefficient, and the value of T is between 0 and 1; n represents the number of points in the track data, m represents the number of moving tracks, SkAnd whether the Kth track has an incidence relation with other tracks or not is shown, the value is 0 or 1, and no incidence relation or incidence relation is shown respectively.
5. The image processing index evaluation method according to claim 3, wherein the error rate is calculated as follows:
Figure FDA0003236726240000021
wherein the content of the first and second substances,
Figure FDA0003236726240000022
Piwhether the ith point in the track data belongs to the track of the same target object or not is represented, and the value is 0 or 1Respectively, do not belong to or belong to; t represents a fault-tolerant coefficient, and the value of T is between 0 and 1; n represents the number of points in the track data, m represents the number of moving tracks, SkAnd whether the Kth track has an incidence relation with other tracks or not is shown, the value is 0 or 1, and no incidence relation or incidence relation is shown respectively.
6. The image processing index evaluation method of claim 3, wherein the leakage rate is calculated as follows:
LRR＝1–RRR
wherein RRR represents a target association rate.
7. The image processing index evaluation method according to claim 3, wherein the attribute accuracy is calculated as follows:
Figure FDA0003236726240000023
wherein the content of the first and second substances,
Figure FDA0003236726240000024
aiwhether the ith point in the track data belongs to the same track data or not is represented, the value is 0 or 1, Tn represents a fault-tolerant coefficient, and the value is between 0 and 1; n represents the number of points in the track data, M represents the number of the track data, and M represents the total number of the track data or the target objects.
8. The image processing index evaluation method of claim 3, wherein the recall ratio is calculated as follows:
RecallR＝PRO/LO
wherein, PRO represents the number of target objects predicted to be correct after the repetition; and LO represents the number of the labeling target objects in the labeling data.
9. The image processing index evaluation method according to claim 3, wherein the repetition rate is calculated as follows:
RepateR＝(PRT-ES)/LO–1
wherein PRT represents the total number of target objects with correct prediction; ES denotes the number of false beats.
10. The image processing index evaluation method according to claim 3, wherein the snapshot accuracy is calculated as follows:
SRR＝(PRT-ES)/PRT
wherein PRT represents the total number of target objects with correct prediction; ES denotes the number of false beats.
11. The image processing index evaluation method according to claim 3, wherein the false capture rate is calculated as follows:
SER＝ES/PRT
wherein PRT represents the total number of target objects with correct prediction; ES denotes the number of false beats.
12. An image processing index evaluation system, comprising:
the system comprises a structuring module, a data processing module and a data processing module, wherein the structuring module is used for acquiring image data, and outputting the structured data after one or more preset algorithm model operations;
the track acquisition module is used for analyzing the structured data and predicting track data corresponding to one or more target objects according to the input marking data; the track acquisition module comprises a label determination unit, and is used for the labeling data to comprise label information of the target object; performing label determination on the structured data according to the label information and a preset label determination algorithm, and obtaining track data corresponding to the same target object based on the determined label information; the preset label determination algorithm comprises a cross ratio method, and the label information comprises the position, ID and/or attribute information of a target object;
and the index evaluation module is used for evaluating the predicted track data and outputting an index evaluation result.
13. The image processing index evaluation system of claim 12, wherein the index evaluation module comprises an algorithm evaluation unit for evaluating the trajectory data according to a preset index evaluation algorithm or algorithms.
14. The image processing index evaluation system of claim 13, wherein the index evaluation result comprises at least one or more of: recall rate, snapshot repetition rate, snapshot accuracy rate, false capture rate, target association rate, false association rate, missed association rate, and attribute accuracy rate.
15. An apparatus, comprising:
one or more processors; and
one or more machine-readable media having instructions stored thereon that, when executed by the one or more processors, cause the apparatus to perform the method recited by one or more of claims 1-11.
16. One or more machine-readable media having instructions stored thereon, which when executed by one or more processors, cause an apparatus to perform the method recited by one or more of claims 1-11.
CN202010879122.8A 2020-08-27 2020-08-27 Image processing index evaluation method, system, equipment and medium Active CN112017171B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010879122.8A CN112017171B (en) 2020-08-27 2020-08-27 Image processing index evaluation method, system, equipment and medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010879122.8A CN112017171B (en) 2020-08-27 2020-08-27 Image processing index evaluation method, system, equipment and medium

Publications (2)

Publication Number Publication Date
CN112017171A CN112017171A (en) 2020-12-01
CN112017171B true CN112017171B (en) 2021-10-26

Family

ID=73502322

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010879122.8A Active CN112017171B (en) 2020-08-27 2020-08-27 Image processing index evaluation method, system, equipment and medium

Country Status (1)

Country Link
CN (1) CN112017171B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113255421A (en) * 2020-12-08 2021-08-13 四川云从天府人工智能科技有限公司 Image detection method, system, device and medium
CN112528938A (en) * 2020-12-22 2021-03-19 四川云从天府人工智能科技有限公司 Vehicle detection model training and detection method, device and computer storage medium thereof
CN112712006A (en) * 2020-12-25 2021-04-27 四川云从天府人工智能科技有限公司 Target picture snapshot method, system, medium and device
CN112733948B (en) * 2021-01-15 2022-09-06 重庆紫光华山智安科技有限公司 Full-target associated index automatic testing method, system, medium and terminal
CN113345037A (en) * 2021-06-07 2021-09-03 重庆紫光华山智安科技有限公司 Automatic testing method, system, medium and terminal for motor vehicle algorithm indexes

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106846374A (en) * 2016-12-21 2017-06-13 大连海事大学 The track calculating method of vehicle under multi-cam scene
CN107545582A (en) * 2017-07-04 2018-01-05 深圳大学 Video multi-target tracking and device based on fuzzy logic
CN107644415A (en) * 2017-09-08 2018-01-30 众安信息技术服务有限公司 A kind of text image method for evaluating quality and equipment
CN109033451A (en) * 2018-08-21 2018-12-18 北京深瞐科技有限公司 People's vehicle dynamic file analysis method and device
CN109444897A (en) * 2018-09-13 2019-03-08 中国船舶重工集团公司第七〇五研究所 A kind of more gusts of Data Associations based on multiple features
CN109885943A (en) * 2019-02-26 2019-06-14 百度在线网络技术（北京）有限公司 Prediction technique, device, storage medium and the terminal device for decision of driving a vehicle
CN109937343A (en) * 2017-06-22 2019-06-25 百度时代网络技术（北京）有限公司 Appraisal framework for the prediction locus in automatic driving vehicle traffic forecast
CN110717414A (en) * 2019-09-24 2020-01-21 青岛海信网络科技股份有限公司 Target detection tracking method, device and equipment

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR3048148B1 (en) * 2016-02-22 2018-12-07 Universite Blaise Pascal LOCATION OF A TARGET FOR FOLLOWING VEHICLE

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN106846374A (en) * 2016-12-21 2017-06-13 大连海事大学 The track calculating method of vehicle under multi-cam scene
CN109937343A (en) * 2017-06-22 2019-06-25 百度时代网络技术（北京）有限公司 Appraisal framework for the prediction locus in automatic driving vehicle traffic forecast
CN107545582A (en) * 2017-07-04 2018-01-05 深圳大学 Video multi-target tracking and device based on fuzzy logic
CN107644415A (en) * 2017-09-08 2018-01-30 众安信息技术服务有限公司 A kind of text image method for evaluating quality and equipment
CN109033451A (en) * 2018-08-21 2018-12-18 北京深瞐科技有限公司 People's vehicle dynamic file analysis method and device
CN109444897A (en) * 2018-09-13 2019-03-08 中国船舶重工集团公司第七〇五研究所 A kind of more gusts of Data Associations based on multiple features
CN109885943A (en) * 2019-02-26 2019-06-14 百度在线网络技术（北京）有限公司 Prediction technique, device, storage medium and the terminal device for decision of driving a vehicle
CN110717414A (en) * 2019-09-24 2020-01-21 青岛海信网络科技股份有限公司 Target detection tracking method, device and equipment

Also Published As

Publication number Publication date
CN112017171A (en) 2020-12-01

Similar Documents

Publication Publication Date Title
CN112017171B (en) Image processing index evaluation method, system, equipment and medium
Henry et al. Multinational license plate recognition using generalized character sequence detection
Singh et al. Deep spatio-temporal representation for detection of road accidents using stacked autoencoder
US20230267735A1 (en) Method for structuring pedestrian information, device, apparatus and storage medium
CN111259868B (en) Reverse vehicle detection method, system and medium based on convolutional neural network
TWI712980B (en) Claim information extraction method and device, and electronic equipment
CN113052079B (en) Regional passenger flow statistical method, system, equipment and medium based on face clustering
CN108986137B (en) Human body tracking method, device and equipment
CN110781806A (en) Pedestrian detection tracking method based on YOLO
CN112766218B (en) Cross-domain pedestrian re-recognition method and device based on asymmetric combined teaching network
CN110796828B (en) Method and system for searching lost population based on Internet of vehicles
CN111340848A (en) Object tracking method, system, device and medium for target area
CN112801235A (en) Model training method, prediction device, re-recognition model and electronic equipment
CN111310725A (en) Object identification method, system, machine readable medium and device
CN112307143A (en) Space-time trajectory construction method, system, device and medium
WO2020103462A1 (en) Video search method and apparatus, computer device, and storage medium
CN111339943A (en) Object management method, system, platform, equipment and medium
CN112686122B (en) Human body and shadow detection method and device, electronic equipment and storage medium
CN111260697A (en) Target object identification method, system, device and medium
Liu et al. Spatio-temporal prediction and reconstruction network for video anomaly detection
CN111753618A (en) Image recognition method and device, computer equipment and computer readable storage medium
CN111818364B (en) Video fusion method, system, device and medium
CN116246299A (en) Low-head-group intelligent recognition system combining target detection and gesture recognition technology
CN115393755A (en) Visual target tracking method, device, equipment and storage medium
Liu et al. Integrated multiscale appearance features and motion information prediction network for anomaly detection

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant