Supercharge Your Innovation With Domain-Expert AI Agents!

Object representation method based on multi-task feature learning

A feature learning and object representation technology, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as difficulty in meeting public security real-time alarm and rapid response, huge amount of video data, and slow event processing speed. The effect of reducing the amount of transmitted data, enriching spatial information, and speeding up the transmission speed

Inactive Publication Date: 2019-10-22
HANGZHOU DIANZI UNIV
View PDF4 Cites 14 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Nevertheless, when major densely populated cities in my country are facing increasingly serious major criminal crimes, public safety hazards, or complex civil disputes, due to the huge amount of video data, various types of data, and low value density, it is difficult to deal with incidents. Serious problems such as too slow speed, it is difficult to meet the needs of real-time alarm and quick response of public security

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
  • Object representation method based on multi-task feature learning
  • Object representation method based on multi-task feature learning
  • Object representation method based on multi-task feature learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0023] The present invention will be described in detail below with reference to the accompanying drawings, and the objects and effects of the present invention will become more apparent.

[0024] The present invention proposes an object representation method based on multi-task feature learning, and its overall network flow chart is as follows figure 1 shown, the specific steps are as follows:

[0025] Step 1: Input the extracted video saliency objects into the two sub-networks of the object feature extraction network respectively, in which the shallow convolutional network has three layers, and each layer performs convolution, batch normalization and activation processing on the object image. In this way, rich spatial information can be obtained while reducing the amount of data; the deep residual network is the resnet101 network, and the deep network performs high-dimensional feature extraction on video objects through operations such as convolution, pooling, and activation...

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 invention discloses an object representation method based on multi-task feature learning. According to the method, the features are extracted at a terminal by adopting a deep neural network, so that the data volume of transmitted videos is reduced, the transmission speed is increased, and the occupied storage space is reduced; the computer analyzes and recovers the object at the cloud end, sothat the investment of labor cost is reduced, and meanwhile, the event processing efficiency is improved; and the effect of simultaneous implementation and joint optimization of a plurality of tasks can be achieved. According to the invention, an artificial intelligence neural network method is adopted, feature extraction is carried out on an original video image, the transmission data volume is reduced, intelligent analysis is realized through feature calculation, image restoration is realized by using a transposed convolution technology, the event processing speed is greatly improved, and funds are saved.

Description

technical field [0001] The invention relates to the fields of computer graphics and multimedia, and in particular to the transmission and analysis of video objects under big data. Background technique [0002] In recent years, with the rapid development of my country's economy, urban public safety construction has become particularly important. Video surveillance cameras have been deployed on a large scale across the country, and their scale is still growing rapidly. Smart security inspection has become an important aspect of public safety prevention. means. According to statistics, my country's "safe city" construction project has built the world's largest five-level security monitoring network covering my country's main streets, districts, cities, provinces and the central government in more than 660 cities across the country. [0003] Nevertheless, when major densely populated cities in my country are facing increasingly severe major criminal crimes, public safety hazards...

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 Applications(China)
IPC IPC(8): G06K9/46G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V10/462G06N3/045G06F18/217G06F18/24
Inventor 颜成钢王廷宇赵崇宇万斌孙垚棋张继勇张勇东蒋云良
Owner HANGZHOU DIANZI UNIV
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More