Traitement en cours

Veuillez attendre...

Paramétrages

Paramétrages

Aller à Demande

1. CN109376696 - Method, apparatus, computer device and storage medium for classifying video actions

Office Chine
Numéro de la demande 201811437221.X
Date de la demande 28.11.2018
Numéro de publication 109376696
Date de publication 22.02.2019
Type de publication A
CIB
G06K 9/00
GPHYSIQUE
06CALCUL; COMPTAGE
KRECONNAISSANCE DES DONNÉES; PRÉSENTATION DES DONNÉES; SUPPORTS D'ENREGISTREMENT; MANIPULATION DES SUPPORTS D'ENREGISTREMENT
9Méthodes ou dispositions pour la lecture ou la reconnaissance de caractères imprimés ou écrits ou pour la reconnaissance de formes, p.ex. d'empreintes digitales
CPC
G06K 9/00718
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
00624Recognising scenes, i.e. recognition of a whole field of perception; recognising scene-specific objects
00711Recognising video content, e.g. extracting audiovisual features from movies, extracting representative key-frames, discriminating news vs. sport content
00718Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
Déposants BEIJING DAJIA INTERCONNECTION INFORMATION TECHNOLOGY CO., LTD.
北京达佳互联信息技术有限公司
Inventeurs ZHANG ZHIWEI
张志伟
LI YAN
李岩
Mandataires 北京三高永信知识产权代理有限责任公司 11138
Titre
(EN) Method, apparatus, computer device and storage medium for classifying video actions
(ZH) 视频动作分类的方法、装置、计算机设备和存储介质
Abrégé
(EN)
The invention relates to a method, an apparatus, a computer device and a storage medium for classifying video actions, belonging to the technical field of machine learning models. The method comprisesthe following steps: obtaining a video to be classified and determining a plurality of video frames in the video to be classified; A plurality of video frames are inputted into an optical flow substitution module in an optimized video motion classification model after training, and optical flow characteristic information corresponding to the plurality of video frames is obtained. A plurality of video frames are input into a three-dimensional convolution neural module in the optimized video motion classification model after training, and the spatial characteristic information corresponding tothe plurality of video frames is obtained. Based on the optical flow characteristic information and the spatial characteristic information, the classification category information corresponding to thevideo to be classified is determined. By adopting the invention, a plurality of video frames of a video to be classified can be directly used as inputs of an optical flow substitution module in the model, and the optical flow substitution module can directly extract optical flow characteristic information corresponding to the plurality of video frames of the video to be classified, thereby further improving the efficiency of classification processing.

(ZH)
本公开是关于一种视频动作分类的方法、装置、计算机设备和存储介质,属于机器学习模型技术领域。所述方法包括:获取待分类视频,确定待分类视频中的多个视频帧;将多个视频帧输入到训练后的优化视频动作分类模型中的光流替代模块中,得到多个视频帧对应的光流特征信息;将多个视频帧输入到训练后的优化视频动作分类模型中的三维卷积神经模块中,得到多个视频帧对应的空间特征信息;基于光流特征信息和空间特征信息,确定待分类视频对应的分类类别信息。采用本公开,可以将待分类视频的多个视频帧直接作为模型中的光流替代模块的输入,光流替代模块可以直接提取待分类视频的多个视频帧对应的光流特征信息,进一步提高了分类处理的效率。

Également publié en tant que