Unlock instant, AI-driven research and patent intelligence for your innovation.

Video classification model training method, classification method, device and equipment

A classification model and video classification technology, applied in the field of computer vision, can solve the problems of high cost and low efficiency of manual labeling, and achieve the effect of high cost, low efficiency and improved prediction accuracy

Active Publication Date: 2022-06-28
GUANGDONG OPPO MOBILE TELECOMM CORP LTD
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The embodiment of the present application provides a training method, classification method, device and equipment for a video classification model, which can solve the problem that although manual labeling has significantly improved the accuracy of labels, the cost of manual labeling is high and the efficiency is low

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Video classification model training method, classification method, device and equipment
  • Video classification model training method, classification method, device and equipment
  • Video classification model training method, classification method, device and equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0042] In order to make the objectives, technical solutions and advantages of the present application clearer, the embodiments of the present application will be further described in detail below with reference to the accompanying drawings.

[0043] First, a brief introduction is made to several technical terms provided by the embodiments of the present application:

[0044] Youtube-8M Video Understanding Challenge: It is a video understanding challenge sponsored by Kaggle and Google, which requires the use of machine learning models under 1G for video label classification tasks. Held once a year, it has been held twice, and 2019 is the third time.

[0045] Youtube-8M dataset: A large labeled dataset with 6.1 million videos and 3862 classes (or labels). The raw video is encoded in this dataset as a series of feature vectors, including visual features and audio features. These features are all frames extracted at 1Hz frequency from the original video, and these features are g...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The application discloses a video classification model training method, classification method, device and equipment. The method includes: obtaining a coarse label data set; obtaining a first classification model and a second classification model, the classification accuracy of the second classification model is higher than that of the first classification model; calling the second classification model to the Predict the soft label of the video in the coarse label data set to obtain a soft label data set, the soft label is to use probability to represent the label of the category of the video; perform the first classification model according to the soft label data set Fine-tuning training to obtain the video classification model. The soft-label data set in this application is generated by machines rather than by manual labeling, which solves the problem of high cost and low efficiency of manual labeling.

Description

technical field [0001] The present application relates to the field of computer vision, and in particular, to a training method, a classification method, an apparatus and equipment for a video classification model. Background technique [0002] Automatically understanding video content has become a key technology for many application scenarios, such as autonomous driving, video-based search, and intelligent robotics. Video tag classification through machine learning is a way to automatically understand video content. [0003] In the related art, a video is encoded into a series of feature vectors, including visual features and audio features, and the feature vectors are input into a trained deep learning model to obtain a label corresponding to the video. This tag is a video-level tag. Typically, the deep learning model is trained based on the Youtube-8M dataset. The Youtube-8M dataset is a large labeled video dataset including 6.1 million video sets and 3862 classes. ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Patents(China)
IPC IPC(8): G06V10/774G06V10/764G06F16/75
CPCG06F16/75G06F18/24G06F18/214
Inventor 尹康
Owner GUANGDONG OPPO MOBILE TELECOMM CORP LTD