WO2018169103A1 - Automatic learning-data generating method and device, and self-directed learning device and method using same - Google Patents

Automatic learning-data generating method and device, and self-directed learning device and method using same Download PDF

Info

Publication number
WO2018169103A1
WO2018169103A1 PCT/KR2017/002784 KR2017002784W WO2018169103A1 WO 2018169103 A1 WO2018169103 A1 WO 2018169103A1 KR 2017002784 W KR2017002784 W KR 2017002784W WO 2018169103 A1 WO2018169103 A1 WO 2018169103A1
Authority
WO
WIPO (PCT)
Prior art keywords
images
image
learning
data
labeled
Prior art date
Application number
PCT/KR2017/002784
Other languages
French (fr)
Korean (ko)
Inventor
김성수
Original Assignee
(주)넥셀
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 (주)넥셀 filed Critical (주)넥셀
Priority to PCT/KR2017/002784 priority Critical patent/WO2018169103A1/en
Publication of WO2018169103A1 publication Critical patent/WO2018169103A1/en

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N99/00—Subject matter not provided for in other groups of this subclass
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/20—Analysis of motion

Definitions

  • the present invention relates to a method and apparatus for generating learning data that can automatically generate labeled learning data required in a machine learning process, and a self learning apparatus and method using the same.
  • Machine learning or machine learning is a field of artificial intelligence, which is a series of processes that train computer-prepared learning data to find an appropriate answer for new input based on the trained knowledge.
  • machine learning or machine learning is a field of artificial intelligence, which is a series of processes that train computer-prepared learning data to find an appropriate answer for new input based on the trained knowledge.
  • training data for training a computer is given both a training input and a training output, the training data is labeled.
  • Object detection is a problem of distinguishing the location and type of a specific object in an image or video.
  • machine learning In order to use machine learning for object detection, a large number of labeled learning data are required.
  • the important thing together with the feature extraction and the learning algorithm is the collection of the labeled learning data, and the more the labeled learning data is provided, the more effective the learning can be performed.
  • an object of the present invention is a self-learning device capable of self-learning using a learning data generation method and apparatus that can automatically generate labeled learning data for machine learning and the labeled learning data generated through the same. And providing a method.
  • Learning data generation method for achieving the above object, receiving a plurality of images having continuity, detecting and tracking the object in the plurality of images, the plurality of images and the detection in each image And storing the recognized object and the recognition rate by tracking, and if there is an image in which the recognition rate of the specific object is greater than or equal to a predetermined threshold value in the plurality of images, using the specific object recognized in the image as the final recognition object. Labeling to generate labeled data.
  • the method may further include detecting a continuous image from an input image and providing the image as the plurality of images, and providing the input image with images collected from a plurality of devices through a network.
  • Self-learning method for achieving the above object, receiving a plurality of images having a continuity, detecting and tracking the object in the plurality of images, the plurality of images and the detection and Storing the recognized object and the recognition rate by tracking; if there is an image in which the recognition rate of a specific object is greater than or equal to a predetermined threshold value in the plurality of images, labeling the plurality of images as a final recognition object by using the specific object recognized in the image Generating and storing labeled data, and machine learning using the labeled data as training data.
  • the apparatus for generating learning data includes an image input unit for receiving a plurality of images having continuity, detecting and tracking objects in the plurality of images, and recognizing objects and recognition rates recognized in each image.
  • An object detection unit for outputting information about the plurality of images, a short-term experience database unit for storing the plurality of images and information about the recognition rate and the object recognized in each image, and the short-term experience database unit to query the plurality of If there is an image in which the recognition rate of the specific object in the image is greater than or equal to a predetermined threshold, the learning data generation unit generates the labeled data by labeling the plurality of images with the specific object recognized in the image as the final recognition object.
  • a self-learning device capable of machine learning using the labeled data generated by using the learning data generation method as training data.
  • labeled learning data for machine learning can be automatically generated.
  • machine learning requiring labeled learning data can be effectively performed, and a self-learning device capable of self-learning using the labeled learning data generated through this can be provided.
  • the image data may be received from various devices distributed and distributed in various places to generate large-scale labeled learning data.
  • FIG. 1 is a block diagram for explaining the configuration of a learning data generating apparatus according to an embodiment of the present invention
  • Figure 2 is a flow chart provided in the description of the learning data collection method according to an embodiment of the present invention.
  • FIG. 4 is a block diagram for explaining a configuration of an apparatus for generating training data according to another embodiment of the present invention.
  • FIG. 5 is a diagram illustrating an example of a self-learning apparatus using the method of generating learning data according to the present invention.
  • a component when referred to as being "connected” or “connected” to another component, a component may be directly connected to or connected to another component, but other components in between It should be understood that the element may exist.
  • Other expressions describing relationships between components, such as “between” or “neighboring”, and the like, such that a component “transmits” a signal to another component, should be interpreted as well.
  • FIG. 1 is a block diagram illustrating a configuration of an apparatus for generating learning data according to an embodiment of the present invention.
  • the apparatus 100 for generating training data may include an image input unit 110, an object detector 120, a short-term experience database unit 150, and a training data generator 160.
  • the object detector 120 may include an object detector 130 and an object tracker 140.
  • Such components may be configured by combining two or more components into one component, or by dividing one or more components into two or more components as necessary when implemented in an actual application.
  • the image input unit 110 may receive image data, and may include a device capable of receiving a camera or other image data.
  • the image input unit 110 may process an image frame such as a still image or a moving image obtained by an image sensor or the like, and output a temporally continuous image including the same object.
  • the object detector 120 may include an object detector 130 and an object tracker 140.
  • the object detector 120 may extract feature points from an image output from the image inputter 110 to determine whether a specific object exists or not. If present, the location and type, and the position of the object in the continuous image can be tracked. As a result, the object detector 120 may output information about the object recognized in each image and the recognition rate of the recognized object.
  • Various methods and algorithms may be used to recognize an object in an image and to calculate recognition rate information of the recognized object.
  • the image output from the image input unit 110, and information about the object and the recognition rate recognized in each image output from the object detector 120 is stored in a database.
  • the training data generation unit 160 inquires the short-term experience database unit 150, and if there is an image having a recognition rate equal to or greater than a preset threshold of a specific object in a plurality of recently stored images, finalize the specific object recognized in the image. Labeled data is generated by labeling a number of recently stored images as recognition objects.
  • FIG. 2 is a flowchart provided to explain an automatic learning data collection method according to an embodiment of the present invention.
  • the object detector 120 executes an object detection and tracking process (S210).
  • the object detection unit 120 detects and tracks the object recognized in each image and the recognition rate for the recognized object in a short-term experience database unit 150 is stored in a database with a plurality of images (S220).
  • the learning data generation unit 160 queries the short-term experience database unit 150 and, when there is an image in which the recognition rate of a specific object is greater than or equal to a preset threshold value in a plurality of images, the specific object recognized in the image as the final recognition object.
  • the recently stored continuity image is labeled to generate labeled data (S240).
  • FIG. 3 is a view referred to for describing the learning data generation method according to the present invention.
  • a portion denoted by reference numeral 300 is a short-term experience with respect to the first image 301, the second image 303, and the third image 305 that are continuous in time including the same object.
  • the object recognized in the first image 301 and the first image and its recognition rate 302 the object recognized in the second image 303 and the second image and its recognition rate 304
  • the object recognized in the third image 305 and the third image and the recognition rate 306 are stored.
  • the first image 301 and the second image 303 may not know clearly which of two objects to be recognized.
  • the portion denoted by reference numeral 320 is determined by the learning data generation unit 160.
  • the first image 301, the second image 303, and the third image 305 are labeled with the final recognition object as an object in which the recognition rate exceeds the threshold value in the three images 305 to generate labeled training data. Indicates the state.
  • the learning data when an object is recognized as a value greater than or equal to a threshold value in one image among consecutive images including the same object, the learning data may be generated by labeling the remaining object as the recognized object. Accordingly, the learning data can be generated based on the image taken from another angle even for an image that is clear but has not been recognized because it was not previously learned.
  • FIG. 4 is a block diagram of an apparatus for generating training data according to another exemplary embodiment.
  • an object detector 420 including an image input unit 410, an object detector 430, and an object tracker 440, a short-term experience database unit 450, and learning data
  • the configuration and function of the generation unit 460 is basically the same as described in the above-described embodiment.
  • the continuity detection unit 470 is further included to detect a part of the continuity in time in which the same object exists in the image input by scene change detection or the like. With such a configuration, it is possible to provide a temporally continuous image in which the same object exists from the input image, and thus, it is possible to generate a labeled learning data by receiving an image from a TV video or various other input sources.
  • an image providing device that can deliver the image data from a plurality of devices, such as a plurality of vehicles or CCTV, distributed in various places through the network, it is also possible to generate large-scale labeled learning data.
  • FIG. 5 shows an example of a self-learning apparatus using the method of generating learning data according to the present invention.
  • an object detector 520 including an image input unit 510, an object detector 530, and an object tracker 540, and a short-term experience database.
  • the configuration and function of the unit 550, the learning data generator 560, and the continuity detector 570 are basically the same as those described in the above-described embodiment.
  • the object detection device 580 is configured to self-learn by using the labeled learning data generated by the learning data generator 560 as training data.
  • the object detection algorithm 580 may update the object recognition algorithm used in the object search unit 520 according to a machine learning object recognition algorithm using the labeled learning data, and may continuously improve performance.
  • the automatically generated labeling training data can be applied to various machine learning.
  • the learning data generation method and apparatus and the self-learning apparatus and method using the same according to the present invention is not limited to the configuration of the embodiments described as described above, the above embodiments can be made in various modifications All or part of each of the embodiments may be configured to be selectively combined so that.
  • the subject matter of the present invention is not limited to the use of hardware or software, and can be applied to any other computing or processing environment. It may be implemented in hardware, software or a combination of hardware software described in the present invention.
  • the present invention can be implemented using circuits. That is, one or more programmable logic circuits, i.e., application specific integrated circuits (ASICs) or logic circuits (AND, OR, NAND gates) or processing devices (e.g., microprocessors, controllers) may be implemented.
  • ASICs application specific integrated circuits
  • AND, OR, NAND gates logic circuits
  • processing devices e.g., microprocessors, controllers
  • the present invention can be implemented as a computer program on a programmable computer.
  • a computer may include a processor, a storage device, an input device, and an output device.
  • Program code may be input by a mouse or a keyboard input device to implement the contents described in the present invention.
  • These programs can be implemented in high-level or object-oriented languages. It can also be implemented as a computer system implemented in assembly or machine code.
  • the present invention can be used to automatically generate labeled learning data for machine learning, and can also be used in a self-learning apparatus and method capable of self-learning using the generated labeled learning data.

Abstract

The present invention relates to a learning-data generating method and device, and a self-directed learning device and method using the same. The learning-data generating method according to the present invention comprises the steps of: receiving multiple images having continuity; detecting and tracking an object in the multiple images; storing the multiple images, an object recognized by the object detection and tracking in each of the images, and the recognition rate of the object; and when, among the multiple images, there is an image in which the recognition rate of a particular object is greater than or equal to a predetermined threshold value, labeling the multiple images by using the particular object recognized in the corresponding image as a final recognized object, so as to generate labeled data. According to the present invention, labeled learning-data required for machine learning can be automatically generated.

Description

자동적인 학습데이터 생성 방법 및 장치와 이를 이용하는 자가 학습 장치 및 방법Automatic learning data generation method and apparatus and self-learning apparatus and method using same
본 발명은 머신 러닝 과정에서 필요한 레이블된 학습데이터를 자동적으로 생성할 수 있는 학습데이터 생성 방법 및 장치와 이를 이용하는 자가 학습 장치 및 방법에 관한 것이다. The present invention relates to a method and apparatus for generating learning data that can automatically generate labeled learning data required in a machine learning process, and a self learning apparatus and method using the same.
머신 러닝(machine learning) 또는 기계 학습은 인공 지능의 한 분야로, 컴퓨터에 미리 준비된 학습데이터를 훈련시켜, 훈련된 지식을 기반으로 새로운 입력에 대하여 적절한 답을 찾고자 하는 일련의 과정이라 할 수 있다. 이때, 컴퓨터를 훈련시키는 학습데이터가 질문(training input)과 정답(training output)이 모두 주어진 경우, 레이블링(labeling) 되어 있다고 한다. Machine learning (machine learning) or machine learning is a field of artificial intelligence, which is a series of processes that train computer-prepared learning data to find an appropriate answer for new input based on the trained knowledge. In this case, when training data for training a computer is given both a training input and a training output, the training data is labeled.
한편, 객체 검출(object detection)은 이미지 또는 동영상에서 특정 객체의 위치와 종류를 구별해내는 문제인데, 객체 검출에 머신 러닝을 이용하기 위해서는, 수많은 레이블된 학습데이터가 필요하다.Object detection, on the other hand, is a problem of distinguishing the location and type of a specific object in an image or video. In order to use machine learning for object detection, a large number of labeled learning data are required.
즉, 머신 러닝을 기반으로 객체 검출을 하는 경우, 특징 추출 및 학습 알고리즘과 함께 중요한 것이 레이블된 학습데이터의 수집에 있으며, 레이블된 학습데이터가 많이 제공되면 될수록, 학습은 더 효과적으로 진행될 수 있다.That is, in the case of object detection based on machine learning, the important thing together with the feature extraction and the learning algorithm is the collection of the labeled learning data, and the more the labeled learning data is provided, the more effective the learning can be performed.
이를 위해서, 수천에서 수만 건의 레이블된 학습데이터가 필요하지만, 레이블된 학습데이터는 일반적으로 수동 작업으로 만들어지고 있는 실정이므로, 방대한 양의 레이블된 학습데이터를 구하는 것은 쉽지 않다. To do this, thousands to tens of thousands of labeled training data are required, but since labeled training data are generally made by manual work, it is not easy to obtain a large amount of labeled training data.
따라서, 객체 검출이나 기타 머신 러닝의 효과적인 학습을 위해서는, 자동적으로 레이블된 학습데이터를 생성하여 이용할 수 있도록 하는 방안이 필요하다.Therefore, for effective learning of object detection or other machine learning, there is a need for a method of automatically generating and using labeled learning data.
따라서, 본 발명의 목적은, 머신 러닝을 위한 레이블된 학습데이터를 자동적으로 생성할 수 있는 학습데이터 생성 방법 및 장치와 이를 통해 생성한 레이블된 학습데이터를 이용하여 자가 학습을 할 수 있는 자가 학습 장치 및 방법을 제공함에 있다.Accordingly, an object of the present invention is a self-learning device capable of self-learning using a learning data generation method and apparatus that can automatically generate labeled learning data for machine learning and the labeled learning data generated through the same. And providing a method.
상기 목적을 달성하기 위한 본 발명에 따른 학습데이터 생성 방법은, 연속성을 갖는 다수의 이미지를 입력받는 단계, 상기 다수의 이미지에서 객체를 탐지 및 추적하는 단계, 상기 다수의 이미지와, 각 이미지에서 탐지 및 추적하여 인식된 객체와 인식률을 저장하는 단계, 및 상기 다수의 이미지에서 특정 객체의 인식률이 기설정된 임계값 이상인 이미지가 있으면, 상기 이미지에서 인식된 특정 객체를 최종 인식 객체로 상기 다수의 이미지를 레이블링하여 레이블된 데이터를 생성하는 단계를 포함한다.Learning data generation method according to the present invention for achieving the above object, receiving a plurality of images having continuity, detecting and tracking the object in the plurality of images, the plurality of images and the detection in each image And storing the recognized object and the recognition rate by tracking, and if there is an image in which the recognition rate of the specific object is greater than or equal to a predetermined threshold value in the plurality of images, using the specific object recognized in the image as the final recognition object. Labeling to generate labeled data.
또한, 입력 영상에서 연속성 있는 이미지를 검출하여, 상기 다수의 이미지로 제공하는 단계와, 네트워크를 통해 다수의 장치로부터 수집한 영상을 상기 입력 영상으로 제공하는 단계를 더 포함할 수 있다. The method may further include detecting a continuous image from an input image and providing the image as the plurality of images, and providing the input image with images collected from a plurality of devices through a network.
상기 목적을 달성하기 위한 본 발명에 따른 자가 학습 방법은, 연속성을 갖는 다수의 이미지를 입력받는 단계, 상기 다수의 이미지에서 객체를 탐지 및 추적하는 단계, 상기 다수의 이미지와, 각 이미지에서 탐지 및 추적하여 인식된 객체와 인식률을 저장하는 단계, 상기 다수의 이미지에서 특정 객체의 인식률이 기설정된 임계값 이상인 이미지가 있으면, 상기 이미지에서 인식된 특정 객체를 최종 인식 객체로 상기 다수의 이미지를 레이블링하여 레이블된 데이터를 생성하여 저장하는 단계, 및 상기 레이블된 데이터를 훈련데이터로 이용하여 기계 학습하는 단계를 포함한다. Self-learning method according to the present invention for achieving the above object, receiving a plurality of images having a continuity, detecting and tracking the object in the plurality of images, the plurality of images and the detection and Storing the recognized object and the recognition rate by tracking; if there is an image in which the recognition rate of a specific object is greater than or equal to a predetermined threshold value in the plurality of images, labeling the plurality of images as a final recognition object by using the specific object recognized in the image Generating and storing labeled data, and machine learning using the labeled data as training data.
또한, 상기 목적을 달성하기 위한 본 발명에 학습데이터 생성 장치는, 연속성을 갖는 다수의 이미지를 입력받는 영상 입력부, 상기 다수의 이미지에서 객체를 탐지 및 추적하여, 각 이미지에서 인식된 객체와 인식률에 대한 정보를 출력하는 객체 탐지부, 상기 다수의 이미지와, 상기 각 이미지에서 인식된 객체와 인식률에 대한 정보를 데이터베이스화하여 저장하는 단기경험 데이터베이스부, 및 상기 단기경험 데이터베이스부를 조회하여, 상기 다수의 이미지에서 특정 객체의 인식률이 기설정된 임계값 이상인 이미지가 있으면, 상기 이미지에서 인식된 특정 객체를 최종 인식 객체로 상기 다수의 이미지를 레이블링하여 레이블된 데이터를 생성하는 학습데이터 생성부를 포함한다.In addition, according to the present invention for achieving the above object, the apparatus for generating learning data includes an image input unit for receiving a plurality of images having continuity, detecting and tracking objects in the plurality of images, and recognizing objects and recognition rates recognized in each image. An object detection unit for outputting information about the plurality of images, a short-term experience database unit for storing the plurality of images and information about the recognition rate and the object recognized in each image, and the short-term experience database unit to query the plurality of If there is an image in which the recognition rate of the specific object in the image is greater than or equal to a predetermined threshold, the learning data generation unit generates the labeled data by labeling the plurality of images with the specific object recognized in the image as the final recognition object.
또한, 상기 목적을 달성하기 위하여 본 발명에서는, 상기 학습데이터 생성 방법을 프로세서에서 실행시키기 위한 프로그램을 기록한 프로세서가 읽을 수 있는 기록매체를 제공할 수 있다. In addition, in order to achieve the above object, in the present invention, it is possible to provide a recording medium that can be read by a processor that records a program for executing the learning data generation method in a processor.
그리고, 본 발명에 따르면, 상기 학습데이터 생성 방법을 이용하여 생성한 레이블된 데이터를 훈련데이터로 이용하여 기계 학습할 수 있는 자가 학습 장치가 제공된다According to the present invention, there is provided a self-learning device capable of machine learning using the labeled data generated by using the learning data generation method as training data.
본 발명에 따르면, 머신 러닝을 위한 레이블된 학습데이터를 자동으로 생성할 수 있다. 이에 의해, 레이블된 학습데이터가 필요한 기계 학습을 효과적으로 진행할 수 있으며, 이를 통해 생성한 레이블된 학습데이터를 이용하여 자가 학습이 가능한 자가 학습 장치도 제공할 수 있다. 또한, 여러 곳에 분산 배치된 다양한 장치로부터 영상 데이터를 입력받아, 대규모로 레이블된 학습데이터를 생성할 수 있다. According to the present invention, labeled learning data for machine learning can be automatically generated. As a result, machine learning requiring labeled learning data can be effectively performed, and a self-learning device capable of self-learning using the labeled learning data generated through this can be provided. In addition, the image data may be received from various devices distributed and distributed in various places to generate large-scale labeled learning data.
도 1은 본 발명의 일실시예에 따른 학습데이터 생성 장치의 구성을 설명하기 위한 블럭도,1 is a block diagram for explaining the configuration of a learning data generating apparatus according to an embodiment of the present invention;
도 2는 본 발명의 일실시예에 따른 학습데이터 수집 방법에 대한 설명에 제공되는 흐름도,Figure 2 is a flow chart provided in the description of the learning data collection method according to an embodiment of the present invention,
도 3은 본 발명의 일실시예에 따른 학습데이터 수집 방법에 대한 설명에 참조되는 도면,3 is a reference to the description of the learning data collection method according to an embodiment of the present invention,
도 4는 본 발명의 다른 실시예에 따른 학습 데이터 생성 장치의 구성을 설명하기 위한 블럭도, 그리고4 is a block diagram for explaining a configuration of an apparatus for generating training data according to another embodiment of the present invention; and
도 5는 본 발명에 따른 학습데이터 생성 방법을 이용하는 자가 학습 장치의 일 예를 나타낸 도면이다. 5 is a diagram illustrating an example of a self-learning apparatus using the method of generating learning data according to the present invention.
본 명세서에서, 어떤 구성요소가 다른 구성요소에 "연결되어" 있다거나 "접속되어" 있다고 언급된 경우, 어떤 구성요소에 다른 구성요소에 직접적으로 연결되어 있거나 접속되어 있을 수도 있지만, 중간에 다른 구성요소가 존재할 수도 있다고 이해되어야 할 것이다. 구성요소들 간의 관계를 설명하는 다른 표현들, 즉 "~사이에" 또는 "~에 이웃하는" 등과, 어떤 구성요소가 다른 구성요소로 신호를 "전송한다" 와 같은 표현도 마찬가지로 해석되어야 한다.In the present specification, when a component is referred to as being "connected" or "connected" to another component, a component may be directly connected to or connected to another component, but other components in between It should be understood that the element may exist. Other expressions describing relationships between components, such as "between" or "neighboring", and the like, such that a component "transmits" a signal to another component, should be interpreted as well.
이하에서는 도면을 참조하여 본 발명을 보다 상세하게 설명한다. Hereinafter, with reference to the drawings will be described the present invention in more detail.
도 1은 본 발명의 일실시예에 따른 학습데이터 생성 장치의 구성을 설명하기 위한 블럭도이다.1 is a block diagram illustrating a configuration of an apparatus for generating learning data according to an embodiment of the present invention.
도 1을 참조하면, 본 학습데이터 생성 장치(100)는 영상 입력부(110), 객체 탐지부(120), 단기경험 데이터베이스부(150), 및 학습데이터 생성부(160)를 포함할 수 있고, 객체 탐지부(120)는 객체 검출부(130)와 객체 추적부(140)를 포함할 수 있다.Referring to FIG. 1, the apparatus 100 for generating training data may include an image input unit 110, an object detector 120, a short-term experience database unit 150, and a training data generator 160. The object detector 120 may include an object detector 130 and an object tracker 140.
이와 같은 구성요소들은 실제 응용에서 구현될 때 필요에 따라 2 이상의 구성요소가 하나의 구성요소로 합쳐지거나, 혹은 하나의 구성요소가 2 이상의 구성요소로 세분되어 구성될 수 있다.Such components may be configured by combining two or more components into one component, or by dividing one or more components into two or more components as necessary when implemented in an actual application.
영상 입력부(110)는 영상 데이터를 입력받으며, 카메라나 기타 영상 데이터를 입력받을 수 있는 장치 등이 포함될 수 있다. 영상 입력부(110)는 이미지 센서 등에 의해 얻어지는 정지영상이나 동영상 등의 화상 프레임을 처리할 수 있으며, 동일한 객체를 포함하는 시간적으로 연속성 있는 이미지를 출력할 수 있다.The image input unit 110 may receive image data, and may include a device capable of receiving a camera or other image data. The image input unit 110 may process an image frame such as a still image or a moving image obtained by an image sensor or the like, and output a temporally continuous image including the same object.
객체 탐지부(120)는 객체 검출부(130)와 객체 추척부(140)를 포함할 수 있으며, 영상 입력부(110)에서 출력되는 이미지에서 특징점 등을 추출하여, 특정 객체가 존재하는지 여부, 객체의 존재시 그 위치와 종류, 연속되는 이미지에서 객체의 위치 추적 등을 할 수 있다. 이에 의해, 객체 탐지부(120)는 각 이미지에서 인식된 객체와 인식된 객체의 인식률에 대한 정보를 출력할 수 있다. 이미지에서 객체의 인식과 인식된 객체에 대한 인식률 정보의 산출에는 다양한 방식과 알고리즘이 사용될 수 있다.The object detector 120 may include an object detector 130 and an object tracker 140. The object detector 120 may extract feature points from an image output from the image inputter 110 to determine whether a specific object exists or not. If present, the location and type, and the position of the object in the continuous image can be tracked. As a result, the object detector 120 may output information about the object recognized in each image and the recognition rate of the recognized object. Various methods and algorithms may be used to recognize an object in an image and to calculate recognition rate information of the recognized object.
단기경험 데이터베이스부(150)에는 영상 입력부(110)에서 출력되는 이미지와, 객체 탐지부(120)에서 출력되는 각 이미지에서 인식된 객체와 인식률에 대한 정보가 데이터베이스화하여 저장된다. In the short-term experience database unit 150, the image output from the image input unit 110, and information about the object and the recognition rate recognized in each image output from the object detector 120 is stored in a database.
학습데이터 생성부(160)는 단기경험 데이터베이스부(150)를 조회하여, 최근 저장된 다수의 이미지에서 특정 객체의 인식률이 미리 설정되어 있는 임계값 이상인 이미지가 있으면, 그 이미지에서 인식된 특정 객체를 최종 인식 객체로 최근 저장된 다수의 이미지를 레이블링하여 레이블된 데이터를 생성한다. The training data generation unit 160 inquires the short-term experience database unit 150, and if there is an image having a recognition rate equal to or greater than a preset threshold of a specific object in a plurality of recently stored images, finalize the specific object recognized in the image. Labeled data is generated by labeling a number of recently stored images as recognition objects.
이와 같은 구성에 의해, 영상 입력부(110)를 통해 입력받은 이미지로부터 레이블된 학습데이터를 자동적으로 생성할 수 있다.By such a configuration, it is possible to automatically generate the learning data labeled from the image received through the image input unit 110.
도 2는 발명의 일실시예에 따른 자동적인 학습데이터 수집 방법에 대한 설명에 제공되는 흐름도이다.2 is a flowchart provided to explain an automatic learning data collection method according to an embodiment of the present invention.
도 2를 참조하면, 영상 입력부(110)를 통해 다수의 이미지가 입력되면(S200), 객체 탐지부(120)에서 객체 검출 및 추적 과정을 실행한다(S210).Referring to FIG. 2, when a plurality of images are input through the image input unit 110 (S200), the object detector 120 executes an object detection and tracking process (S210).
객체 탐지부(120)에서 검출 및 추적하여 각 이미지에서 인식된 객체와 인식된 객체에 대한 인식률은 단기경험 데이터베이스부(150)에 다수의 이미지와 함께 데이터베이스화되어 저장된다(S220).The object detection unit 120 detects and tracks the object recognized in each image and the recognition rate for the recognized object in a short-term experience database unit 150 is stored in a database with a plurality of images (S220).
학습데이터 생성부(160)는 단기경험 데이터베이스부(150)를 조회하여, 다수의 이미지에서 특정 객체의 인식률이 미리 설정된 임계값 이상인 이미지가 있는 경우, 그 이미지에서 인식된 특정 객체를 최종 인식 객체로 최근 저장된 연속성 있는 이미지를 레이블링하여 레이블된 데이터를 생성한다(S240).The learning data generation unit 160 queries the short-term experience database unit 150 and, when there is an image in which the recognition rate of a specific object is greater than or equal to a preset threshold value in a plurality of images, the specific object recognized in the image as the final recognition object. The recently stored continuity image is labeled to generate labeled data (S240).
이와 같은 과정에 의해, 다수의 이미지에 대하여 자동적으로 레이블된 학습데이터를 생성할 수 있다.By this process, it is possible to automatically generate the learning data labeled for a plurality of images.
도 3은 본 발명에 따른 학습데이터 생성 방법에 대한 설명에 참조되는 도면이다. 3 is a view referred to for describing the learning data generation method according to the present invention.
도 3을 참조하면, 도면부호 300으로 표기한 부분이, 동일한 객체를 포함하여 시간적으로 연속적인 제1 이미지(301), 제2 이미지(303), 및 제3 이미지(305)에 대하여, 단기경험 데이터베이스부(150)에, 제1 이미지(301)와 제1 이미지에서 인식된 객체와 그 인식률(302), 제2 이미지(303)와 제2 이미지에서 인식된 객체와 그 인식률(304), 제3 이미지(305)와 제3 이미지에서 인식된 객체와 그 인식률(306)이 저장된 상태를 나타낸다. Referring to FIG. 3, a portion denoted by reference numeral 300 is a short-term experience with respect to the first image 301, the second image 303, and the third image 305 that are continuous in time including the same object. In the database unit 150, the object recognized in the first image 301 and the first image and its recognition rate 302, the object recognized in the second image 303 and the second image and its recognition rate 304, The object recognized in the third image 305 and the third image and the recognition rate 306 are stored.
이와 같은 검출 결과, 제1 이미지(301)와 제2 이미지(303)에서는 두 개의 인식 개체 중 어느 객체로 인식되어야 하는지 명확하게 알 수 없는 상태이다. 그러나, 제3 이미지(305)에서 어느 한 객체의 인식률이 99%인 경우로, 미리 설정된 임계값을 넘는다고 가정할 경우, 도면부호 320으로 표기한 부분은, 학습데이터 생성부(160)에서 제3 이미지(305)에서 인식률이 임계값을 넘는 객체를 최종 인식 객체로 제1 이미지(301), 제2 이미지(303), 및 제3 이미지(305)를 레이블링하여, 레이블된 학습데이터를 생성한 상태를 나타낸다. As a result of the detection, the first image 301 and the second image 303 may not know clearly which of two objects to be recognized. However, when the recognition rate of any one object in the third image 305 is 99% and it is assumed that the predetermined threshold value is exceeded, the portion denoted by reference numeral 320 is determined by the learning data generation unit 160. The first image 301, the second image 303, and the third image 305 are labeled with the final recognition object as an object in which the recognition rate exceeds the threshold value in the three images 305 to generate labeled training data. Indicates the state.
이와 같이 동일한 객체를 포함하며 시간적으로 연속적인 이미지 중에서 어느 한 이미지에서 객체가 임계값 이상으로 인식되면, 나머지 이미지에 대해서도 인식된 객체로 레이블링하여 학습데이터를 생성할 수 있다. 따라서, 이미지는 명확하지만 기존에 학습되지 않아서 인식되지 않았던 이미지에 대해서도 다른 각도에서 촬영한 이미지 등을 기반으로 학습데이터를 생성할 수 있다.As described above, when an object is recognized as a value greater than or equal to a threshold value in one image among consecutive images including the same object, the learning data may be generated by labeling the remaining object as the recognized object. Accordingly, the learning data can be generated based on the image taken from another angle even for an image that is clear but has not been recognized because it was not previously learned.
도 4는 다른 실시예에 따른 학습데이터 생성 장치에 대한 블럭도이다.4 is a block diagram of an apparatus for generating training data according to another exemplary embodiment.
도 4를 참조하면, 본 실시예에서, 영상 입력부(410), 객체 검출부(430)와 객체 추적부(440)를 포함하는 객체 탐지부(420), 단기경험 데이터베이스부(450), 및 학습데이터 생성부(460)의 구성 및 기능은 기본적으로 전술한 실시예에서 설명한 바와 동일하다.Referring to FIG. 4, in the present embodiment, an object detector 420 including an image input unit 410, an object detector 430, and an object tracker 440, a short-term experience database unit 450, and learning data The configuration and function of the generation unit 460 is basically the same as described in the above-described embodiment.
다만, 본 실시예에서는, 연속성 검출부(470)를 더 포함하여, 장면 전환 검출 등에 의해 입력되는 영상에서 동일 객체가 존재하는 시간적으로 연속성이 있는 부분을 검출할 수 있도록 한다. 이와 같은 구성에 의해, 입력 영상으로부터 동일 객체가 존재하는 시간적으로 연속하는 이미지를 제공할 수 있으므로, TV 동영상이나 기타 다양한 입력 소소로부터 이미지를 입력받아서 레이블된 학습데이터를 생성할 수 있다. However, in the present embodiment, the continuity detection unit 470 is further included to detect a part of the continuity in time in which the same object exists in the image input by scene change detection or the like. With such a configuration, it is possible to provide a temporally continuous image in which the same object exists from the input image, and thus, it is possible to generate a labeled learning data by receiving an image from a TV video or various other input sources.
또한, 네트워크를 통해 다수의 차량이나 CCTV 등 여러 곳에 분산되어 있는 다수의 장치로부터 영상 데이터를 전달할 수 있는 영상 제공 장치를 설치하여, 대규모로 레이블된 학습데이터를 생성할 수도 있다.In addition, by installing an image providing device that can deliver the image data from a plurality of devices, such as a plurality of vehicles or CCTV, distributed in various places through the network, it is also possible to generate large-scale labeled learning data.
도 5는 본 발명에 따른 학습데이터 생성 방법을 이용하는 자가 학습 장치의 일 예를 나타낸 것이다.5 shows an example of a self-learning apparatus using the method of generating learning data according to the present invention.
도 5를 참조하면, 본 실시예에 따른 자가 학습 장치(500)에서, 영상 입력부(510), 객체 검출부(530)와 객체 추적부(540)를 포함하는 객체 탐지부(520), 단기경험 데이터베이스부(550), 학습데이터 생성부(560), 및 연속성 검출부(570)의 구성 및 기능은 기본적으로 전술한 실시예에서 설명한 바와 동일하다.Referring to FIG. 5, in the self-learning apparatus 500 according to the present embodiment, an object detector 520 including an image input unit 510, an object detector 530, and an object tracker 540, and a short-term experience database. The configuration and function of the unit 550, the learning data generator 560, and the continuity detector 570 are basically the same as those described in the above-described embodiment.
다만, 본 실시예에서는 물체 검출 장치(580)가 학습데이터 생성부(560)에서 생성한 레이블된 학습데이터를 훈련데이터로 이용하여 스스로 자가 학습할 수 있도록 구성된다.However, in the present embodiment, the object detection device 580 is configured to self-learn by using the labeled learning data generated by the learning data generator 560 as training data.
또한, 물체 검출 장치(580)에서 레이블된 학습데이터를 이용하여 기계 학습된 객체 인식 알고리즘 등에 따라, 객체 탐색부(520)에서 사용되는 객체 인식 알고리즘 등을 업데이트하여 성능을 지속적으로 향상시킬 수 있다. In addition, the object detection algorithm 580 may update the object recognition algorithm used in the object search unit 520 according to a machine learning object recognition algorithm using the labeled learning data, and may continuously improve performance.
이와 같이 자동적으로 생성한 레이블된 학습데이터를 이용하여 다양한 기계 학습에 응용할 수 있다. The automatically generated labeling training data can be applied to various machine learning.
한편, 본 발명에 따른 학습데이터 생성 방법 및 장치와 이를 이용한 자가 학습 장치 및 방법은 상기한 바와 같이 설명된 실시예들의 구성에 한정되게 적용될 수 있는 것이 아니라, 상기한 실시예들은 다양한 변형이 이루어질 수 있도록 각 실시예들의 전부 또는 일부가 선택적으로 조합되어 구성될 수도 있다.On the other hand, the learning data generation method and apparatus and the self-learning apparatus and method using the same according to the present invention is not limited to the configuration of the embodiments described as described above, the above embodiments can be made in various modifications All or part of each of the embodiments may be configured to be selectively combined so that.
그리고, 본 발명의 내용은 하드웨어나 소프트웨어 사용에만 국한되지는 않으며, 다른 어떤 컴퓨팅 또는 처리 환경에 대해서도 적용 가능하다. 본 발명에서 설명하는 하드웨어, 소프트웨어 또는 하드웨어소프트웨어의 조합으로 구현될 수 있다. 본 발명은 회로를 사용하여 구현될 수 있다. 즉, 한 개 이상의 프로그램 가능한 논리회로, 즉 ASIC(Application Specific Integrated Circuit) 또는 논리회로(AND, OR, NAND gates) 또는 프로세싱 장치(예를 들면 마이크로 프로세서, 컨트롤러)로 구현 가능하다. In addition, the subject matter of the present invention is not limited to the use of hardware or software, and can be applied to any other computing or processing environment. It may be implemented in hardware, software or a combination of hardware software described in the present invention. The present invention can be implemented using circuits. That is, one or more programmable logic circuits, i.e., application specific integrated circuits (ASICs) or logic circuits (AND, OR, NAND gates) or processing devices (e.g., microprocessors, controllers) may be implemented.
또한, 본 발명은 프로그램 가능한 컴퓨터상에서 컴퓨터 프로그램으로 구현 가능하다. 이러한 컴퓨터는 프로세서, 저장장치, 입력장치, 출력 장치를 포함할 수 있다. 본 발명에서 설명한 내용을 구현하기 위해 프로그램 코드는 마우스 또는 키보드 입력장치로 입력될 수 있다. 이러한 프로그램들은 고차원적인 언어나, 객체지향적인 언어로 구현될 수 있다. 또한 어셈블리나 기계어 코드로 구현된 컴퓨터 시스템으로도 구현될 수 있다. In addition, the present invention can be implemented as a computer program on a programmable computer. Such a computer may include a processor, a storage device, an input device, and an output device. Program code may be input by a mouse or a keyboard input device to implement the contents described in the present invention. These programs can be implemented in high-level or object-oriented languages. It can also be implemented as a computer system implemented in assembly or machine code.
또한, 이상에서는 본 발명의 바람직한 실시예에 대하여 도시하고 설명하였지만, 본 발명은 상술한 특정의 실시예에 한정되지 아니하며, 청구범위에서 청구하는 본 발명의 요지를 벗어남이 없이 당해 발명이 속하는 기술분야에서 통상의 지식을 가진자에 의해 다양한 변형실시가 가능한 것은 물론이고, 이러한 변형실시들은 본 발명의 기술적 사상이나 전망으로부터 개별적으로 이해되어서는 안될 것이다.In addition, although the preferred embodiment of the present invention has been shown and described above, the present invention is not limited to the specific embodiments described above, but the technical field to which the invention belongs without departing from the spirit of the invention claimed in the claims. Of course, various modifications can be made by those skilled in the art, and these modifications should not be individually understood from the technical spirit or the prospect of the present invention.
본 발명은 머신 러닝을 위한 레이블된 학습데이터를 자동적으로 생성하는데 활용할 수 있으며, 이를 통해 생성한 레이블된 학습데이터를 이용하여 자가 학습을 할 수 있는 자가 학습 장치 및 방법에도 활용할 수 있다.The present invention can be used to automatically generate labeled learning data for machine learning, and can also be used in a self-learning apparatus and method capable of self-learning using the generated labeled learning data.

Claims (15)

  1. 연속성을 갖는 다수의 이미지를 입력받는 단계;Receiving a plurality of images having continuity;
    상기 다수의 이미지에서 객체를 탐지 및 추적하는 단계;Detecting and tracking an object in the plurality of images;
    상기 다수의 이미지와, 각 이미지에서 탐지 및 추적하여 인식된 객체와 인식률을 저장하는 단계; 및Storing the plurality of images, objects detected and tracked in each image, and recognition rates; And
    상기 다수의 이미지에서 특정 객체의 인식률이 기설정된 임계값 이상인 이미지가 있으면, 상기 이미지에서 인식된 특정 객체를 최종 인식 객체로 상기 다수의 이미지를 레이블링하여 레이블된 데이터를 생성하는 단계를 포함하는 학습데이터 생성 방법.If there is an image in which the recognition rate of a specific object is greater than or equal to a predetermined threshold value in the plurality of images, learning data comprising labeling the plurality of images with a specific object recognized in the image as a final recognition object to generate labeled data; How to produce.
  2. 제1항에 있어서,The method of claim 1,
    입력 영상에서 연속성 있는 이미지를 검출하여, 상기 다수의 이미지로 제공하는 단계를 더 포함하는 학습데이터 생성 방법.And detecting a continuous image from an input image and providing the image as the plurality of images.
  3. 제2항에 있어서,The method of claim 2,
    네트워크를 통해 다수의 장치로부터 수집한 영상을 상기 입력 영상으로 제공하는 단계를 더 포함하는 학습데이터 생성 방법.And providing the images collected from the plurality of devices through the network as the input images.
  4. 제1항에 있어서, The method of claim 1,
    상기 레이블된 데이터를 기계 학습을 위한 훈련데이터로 제공하는 단계를 더 포함하는 학습데이터 생성 방법.The training data generation method further comprising the step of providing the labeled data as training data for machine learning.
  5. 제4항에 있어서,The method of claim 4, wherein
    상기 레이블된 데이터를 훈련데이터로 이용하여 기계 학습된 장치로부터 피드백된 알고리즘에 따라, 상기 다수의 이미지에서 객체를 탐지 및 추적하는 알고리즘을 업데이트 하는 단계를 더 포함하는 학습데이터 생성 방법. Updating the algorithm for detecting and tracking objects in the plurality of images according to an algorithm fed back from a machine-learned device using the labeled data as training data.
  6. 연속성을 갖는 다수의 이미지를 입력받는 단계;Receiving a plurality of images having continuity;
    상기 다수의 이미지에서 객체를 탐지 및 추적하는 단계;Detecting and tracking an object in the plurality of images;
    상기 다수의 이미지와, 각 이미지에서 탐지 및 추적하여 인식된 객체와 인식률을 저장하는 단계; Storing the plurality of images, objects detected and tracked in each image, and recognition rates;
    상기 다수의 이미지에서 특정 객체의 인식률이 기설정된 임계값 이상인 이미지가 있으면, 상기 이미지에서 인식된 특정 객체를 최종 인식 객체로 상기 다수의 이미지를 레이블링하여 레이블된 데이터를 생성하여 저장하는 단계; 및If there is an image in which the recognition rate of a specific object is greater than or equal to a predetermined threshold value in the plurality of images, labeling the plurality of images as a final recognition object using the specific object recognized in the image to generate and store labeled data; And
    상기 레이블된 데이터를 훈련데이터로 이용하여 기계 학습하는 단계를 포함하는 자가 학습 방법.Self-learning method using the labeled data as training data.
  7. 제6항에 있어서,The method of claim 6,
    네트워크를 통해 다수의 장치로부터 수집한 영상에서 연속성 있는 이미지를 검출하여, 상기 다수의 이미지로 제공하는 단계를 더 포함하는 자가 학습 방법.And detecting a continuous image from the images collected from the plurality of devices through a network, and providing the continuous image as the plurality of images.
  8. 연속성을 갖는 다수의 이미지를 입력받는 영상 입력부;An image input unit configured to receive a plurality of images having continuity;
    상기 다수의 이미지에서 객체를 탐지 및 추적하여, 각 이미지에서 인식된 객체와 인식률에 대한 정보를 출력하는 객체 탐지부;An object detector for detecting and tracking objects in the plurality of images and outputting information on the objects and recognition rates recognized in each image;
    상기 다수의 이미지와, 상기 각 이미지에서 인식된 객체와 인식률에 대한 정보를 데이터베이스화하여 저장하는 단기경험 데이터베이스부; 및A short-term experience database unit for storing the plurality of images, information on objects recognized in each image, and recognition rate in a database; And
    상기 단기경험 데이터베이스부를 조회하여, 상기 다수의 이미지에서 특정 객체의 인식률이 기설정된 임계값 이상인 이미지가 있으면, 상기 이미지에서 인식된 특정 객체를 최종 인식 객체로 상기 다수의 이미지를 레이블링하여 레이블된 데이터를 생성하는 학습데이터 생성부를 포함하는 학습데이터 생성 장치.Inquiring the short-term experience database unit, if there is an image in which the recognition rate of a specific object in the plurality of images is greater than or equal to a predetermined threshold value, the specific object recognized in the image is labeled as a final recognition object to label the plurality of images. Learning data generation device comprising a learning data generation unit for generating.
  9. 제8항에 있어서,The method of claim 8,
    입력 영상에서 연속성 있는 이미지를 검출하여, 상기 다수의 이미지로 제공하는 연속성 검출부를 더 포함하는 학습데이터 생성 장치.And a continuity detector configured to detect continuity images from an input image and provide the plural images as the plurality of images.
  10. 제8항에 있어서,The method of claim 8,
    네트워크를 통해 다수의 장치로부터 수집한 영상을 상기 입력 영상으로 제공하는 영상 제공 장치를 더 포함하는 학습데이터 생성 장치.And a video providing device configured to provide, as the input video, images collected from a plurality of devices through a network.
  11. 제8항에 있어서,The method of claim 8,
    상기 레이블된 데이터를 이용하여 기계 학습되는 물체 검출 장치를 더 포함하는 학습데이터 생성 장치.And an object detection device which is machine-learned using the labeled data.
  12. 제11항에 있어서,The method of claim 11,
    상기 객체 탐지부는, 상기 물체 검출 장치에서 학습된 객체 인식 알고리즘에 따라, 인식 알고리즘을 업데이트 하는 것을 특징으로 하는 학습데이터 생성 장치.And the object detecting unit updates the recognition algorithm according to the object recognition algorithm learned by the object detecting device.
  13. 제1항 내지 제5항 중 어느 한 항의 학습데이터 생성 방법을 프로세서에서 실행시키기 위한 프로그램을 기록한 프로세서가 읽을 수 있는 기록매체. A processor-readable recording medium having recorded thereon a program for executing the method of generating learning data according to claim 1.
  14. 제6항 또는 제7항의 자가 학습 방법을 프로세서에서 실행시키기 위한 프로그램을 기록한 프로세서가 읽을 수 있는 기록매체. A processor-readable recording medium having recorded thereon a program for executing the self-learning method of claim 6 or 7.
  15. 제1항 내지 제5항 중 어느 한 항의 학습데이터 생성 방법을 이용하여 생성한 레이블된 데이터를 훈련데이터로 이용하여 기계 학습하는 자가 학습 장치.A self-learning apparatus for machine learning using the labeled data generated by using the method of generating training data of claim 1 as training data.
PCT/KR2017/002784 2017-03-15 2017-03-15 Automatic learning-data generating method and device, and self-directed learning device and method using same WO2018169103A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/KR2017/002784 WO2018169103A1 (en) 2017-03-15 2017-03-15 Automatic learning-data generating method and device, and self-directed learning device and method using same

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/KR2017/002784 WO2018169103A1 (en) 2017-03-15 2017-03-15 Automatic learning-data generating method and device, and self-directed learning device and method using same

Publications (1)

Publication Number Publication Date
WO2018169103A1 true WO2018169103A1 (en) 2018-09-20

Family

ID=63522501

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2017/002784 WO2018169103A1 (en) 2017-03-15 2017-03-15 Automatic learning-data generating method and device, and self-directed learning device and method using same

Country Status (1)

Country Link
WO (1) WO2018169103A1 (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20130132311A1 (en) * 2011-11-18 2013-05-23 Honeywell International Inc. Score fusion and training data recycling for video classification
US20130142418A1 (en) * 2011-12-06 2013-06-06 Roelof van Zwol Ranking and selecting representative video images
US8762299B1 (en) * 2011-06-27 2014-06-24 Google Inc. Customized predictive analytical model training
US20140279739A1 (en) * 2013-03-15 2014-09-18 InsideSales.com, Inc. Resolving and merging duplicate records using machine learning
US20150033362A1 (en) * 2012-02-03 2015-01-29 See-Out Pty Ltd. Notification and Privacy Management of Online Photos and Videos

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8762299B1 (en) * 2011-06-27 2014-06-24 Google Inc. Customized predictive analytical model training
US20130132311A1 (en) * 2011-11-18 2013-05-23 Honeywell International Inc. Score fusion and training data recycling for video classification
US20130142418A1 (en) * 2011-12-06 2013-06-06 Roelof van Zwol Ranking and selecting representative video images
US20150033362A1 (en) * 2012-02-03 2015-01-29 See-Out Pty Ltd. Notification and Privacy Management of Online Photos and Videos
US20140279739A1 (en) * 2013-03-15 2014-09-18 InsideSales.com, Inc. Resolving and merging duplicate records using machine learning

Similar Documents

Publication Publication Date Title
EP3673417B1 (en) System and method for distributive training and weight distribution in a neural network
WO2019107614A1 (en) Machine vision-based quality inspection method and system utilizing deep learning in manufacturing process
WO2017164478A1 (en) Method and apparatus for recognizing micro-expressions through deep learning analysis of micro-facial dynamics
CN107886048A (en) Method for tracking target and system, storage medium and electric terminal
WO2019040213A1 (en) Adaptive real-time detection and examination network (arden)
WO2015178540A1 (en) Apparatus and method for tracking target using handover between cameras
WO2016013885A1 (en) Method for retrieving image and electronic device thereof
WO2021107650A1 (en) Jointly learning visual motion and confidence from local patches in event cameras
US20140226857A1 (en) Information processing system, information processing method and program
WO2020246655A1 (en) Situation recognition method and device for implementing same
WO2021075772A1 (en) Object detection method and device using multiple area detection
WO2021020866A1 (en) Image analysis system and method for remote monitoring
WO2021101045A1 (en) Electronic apparatus and method for controlling thereof
WO2022059955A1 (en) Posture determination system based on radar point clouds
WO2018111011A1 (en) Moving object detection system and method
CN112307864A (en) Method and device for determining target object and man-machine interaction system
WO2012137994A1 (en) Image recognition device and image-monitoring method therefor
Harish et al. Intruder detection by extracting semantic content from surveillance videos
WO2022197136A1 (en) System and method for enhancing machine learning model for audio/video understanding using gated multi-level attention and temporal adversarial training
CN114359976A (en) Intelligent security method and device based on person identification
WO2018169103A1 (en) Automatic learning-data generating method and device, and self-directed learning device and method using same
CN114140751B (en) Examination room monitoring method and system
US11722763B2 (en) System and method for audio tagging of an object of interest
WO2022019601A1 (en) Extraction of feature point of object from image and image search system and method using same
CN109815921A (en) The prediction technique and device of the class of activity in hydrogenation stations

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: 17901126

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 030120)

122 Ep: pct application non-entry in european phase

Ref document number: 17901126

Country of ref document: EP

Kind code of ref document: A1