Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Image content information analysis method based on characteristic variable algorithm

A technology of image content and information analysis, applied in computing, computer components, instruments, etc., can solve problems such as not being able to meet the detection requirements of vulgar image content, improve the overall recognition accuracy, increase the distinction between categories, and increase awareness effect of ability

Inactive Publication Date: 2017-05-03
TAIJI COMP
View PDF0 Cites 15 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Simple training methods and result processing modes cannot meet the needs of various vulgar image content detection. Therefore, it is of great research value to study a hierarchical classification method and result optimization strategy suitable for deep network model vulgar content in the field of vulgar content detection. and application prospects

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
  • Image content information analysis method based on characteristic variable algorithm
  • Image content information analysis method based on characteristic variable algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0017] Considering that there are a large number of vulgar and restricted content pictures in the current network that cannot be accurately and quickly detected and filtered, this invention uses a deep learning network model to identify and classify picture content, and introduces hierarchical classification and result optimization strategies to optimize the network model , greatly improving the detection accuracy.

[0018] Such as figure 1 As shown, a method for analyzing image content information based on a characteristic variable algorithm of the present invention, before using data to train the deep network model, first performs hierarchical classification processing on the training sample set, and the specific steps are as follows:

[0019] (1) Divide all training sample set pictures into multiple first-level categories according to requirements, and divide them into multiple first-level categories on the basis of the first-level categories;

[0020] (2) Use the trained ...

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 relates to the technical field of image content information analysis, particularly to an image content information analysis method based on a characteristic variable algorithm. The method comprises the following steps: (1) dividing all training sample set pictures into a plurality of primary big classes according to demands, and secondarily dividing the primary big classes into a plurality of primary classes; (2) classifying the pictures by using a trained deep network model, calculating the model to obtain the confidence Pi of each class, comparing the confidence difference P=PTOP1-PTOP2 with a threshold Th, and if P is smaller than the threshold Th, performing corresponding optimization strategy adjustment according to the PTOP1 and PTOP2 classes; and if P is greater than the threshold Th, considering the classification result confidential, and directly outputting the classification result without adjusting. The method adopts a multi-level classification strategy, thereby increasing the cognitive ability of a deep learning network on picture characteristics, and improving the accuracy of overall recognition.

Description

technical field [0001] The invention relates to the technical field of image content information analysis, in particular to an image content information analysis method based on a characteristic variable algorithm. Background technique [0002] With the development of Internet technology, the speed of information dissemination is getting faster and faster. While a large amount of information can be obtained, the vulgar content and pictures wantonly spread on the Internet pollute the network environment. How to quickly and accurately identify the pictures containing vulgar content is an urgent problem. The problem. The current methods for detecting vulgar adults in pictures are generally divided into two categories. One is manual detection. This method requires high labor costs, slow detection speed, and inconsistent standards. For a large amount of Internet information, the effect is not ideal; the other method is simple discrimination based on the color of the entire image...

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
IPC IPC(8): G06K9/62
CPCG06F18/24G06F18/214
Inventor 李慧
Owner TAIJI COMP
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products