Processing

Please wait...

Settings

Settings

Goto Application

1. CN111768432 - Moving target segmentation method and system based on twin deep neural network

Office
China
Application Number 202010619639.3
Application Date 30.06.2020
Publication Number 111768432
Publication Date 13.10.2020
Grant Number 111768432
Grant Date 10.06.2022
Publication Kind B
IPC
G06T 7/246
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
7Image analysis
20Analysis of motion
246using feature-based methods, e.g. the tracking of corners or segments
G06T 7/194
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
7Image analysis
10Segmentation; Edge detection
194involving foreground-background segmentation
G06K 9/62
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
G06N 3/08
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
G06N 3/04
GPHYSICS
06COMPUTING; CALCULATING OR COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architecture, e.g. interconnection topology
G06V 10/80
CPC
G06T 7/246
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
7Image analysis
20Analysis of motion
246using feature-based methods, e.g. the tracking of corners or segments
G06T 7/194
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
TIMAGE DATA PROCESSING OR GENERATION, IN GENERAL
7Image analysis
10Segmentation; Edge detection
194involving foreground-background segmentation
G06K 9/629
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
KRECOGNITION OF DATA; PRESENTATION OF DATA; RECORD CARRIERS; HANDLING RECORD CARRIERS
9Methods or arrangements for reading or recognising printed or written characters or for recognising patterns, e.g. fingerprints
62Methods or arrangements for recognition using electronic means
6288Fusion techniques, i.e. combining data from various sources, e.g. sensor fusion
629of extracted features
G06N 3/08
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
08Learning methods
G06N 3/0454
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architectures, e.g. interconnection topology
0454using a combination of multiple neural nets
G06N 3/0445
GPHYSICS
06COMPUTING; CALCULATING; COUNTING
NCOMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS
3Computer systems based on biological models
02using neural network models
04Architectures, e.g. interconnection topology
0445Feedback networks, e.g. hopfield nets, associative networks
Applicants INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES
中国科学院自动化研究所
INSTITUTE OF AUTOMATION, CHINESE ACADEMY OF SCIENCES, GUANGZHOU AI AND ADVANCED COMPUTING INSTITUTE
广东人工智能与先进计算研究院
Inventors ZOU ZHUOJUN
邹卓君
HAO JIE
蒿杰
SHU LIN
舒琳
LIANG JUN
梁俊
GUO YAO
郭尧
Agents 北京市恒有知识产权代理事务所(普通合伙) 11576
北京市恒有知识产权代理事务所(普通合伙) 11576
Title
(EN) Moving target segmentation method and system based on twin deep neural network
(ZH) 基于孪生深度神经网络的动目标分割方法及系统
Abstract
(EN) The invention relates to a moving target segmentation method and system based on a twin deep neural network, and the method comprises the steps: obtaining a plurality of groups of historical image information, wherein each group of historical image information comprises a current frame and a reference frame which are the same in size in the same video, and a label marked with the movement condition of a target; training a VGG16 network model according to each group of historical image information; and according to the trained VGG16 network model, carrying out motion transformation detection and/or relative background detection on a to-be-detected image, and determining the moving target condition in the to-be-detected image. According to the invention, multiple groups of current frames, reference frames and labels are used to train a VGG16 network model, time dimension information is compared with template frames, and templates are flexibly selected in a twin network, so that the method can well adapt to motion photography conditions under the condition of utilizing the time dimension information, and the accuracy of moving target segmentation is effectively improved.
(ZH) 本发明涉及一种基于孪生深度神经网络的动目标分割方法及系统,所述分割方法包括:获取多组历史图像信息,每组历史图像信息包括同一视频中、尺寸大小相同的当前帧和参考帧、以及标有目标的运动情况的标签;根据各组历史图像信息,训练VGG16网络模型;根据训练后的VGG16网络模型,对待检测图像进行运动变换检测和相对背景检测,确定所述待检测图像中的动目标情况。本发明通过多组当前帧、参考帧及标签,对VGG16网络模型训练,将时间维度的信息通过对模板帧的对比,由于孪生网络中对模板的灵活选取,使本发明能够在利用时间维度信息的情况下良好地适应运动摄影条件,有效提高对动目标分割的准确度。
Related patent documents