Method for designing manifold based regularization based semi-supervised classifier for dynamic vision

A design method and classifier technology, applied in the direction of instruments, calculations, computer components, etc., can solve the problems of not being able to guarantee the sparsity of the classifier, affecting the speed of the classifier, etc.

CN102129570AActive Publication Date: 2011-07-20INST OF AUTOMATION CHINESE ACAD OF SCI
0 Cites 9 Cited by

Patent Information

Authority / Receiving Office
CN · China
Current Assignee / Owner
Publication Date
2011-07-20

Smart Images

  • Figure 1
    Figure 1
  • Figure 2
    Figure 2
  • Figure 3
    Figure 3
Patent Text Reader

Abstract

The invention discloses a method for designing a manifold regularization based semi-supervised classifier for dynamic vision. A training data source of the classifier under the environment of the dynamic vision is as follows: data with class information acquired by a user, positive class sample including objects, negative class sample not including objects, and data without class information randomly acquired by a computer in a video; continuity regular terms of the classifier on a sample data aggregation is defined by utilizing a local linear reconstruction coefficient of the data so as to make the utilization rate of the classifier on the data without the class information be improved. In addition, the definitions of a function complexity and function continuity regular terms are configured to be a form of 1 norm in an optimization problem of solving the coefficients of the classifier, thus the solution of the optimization problem of the classifier is a sparse solution of the coefficients of the classifier, namely the classifier obtained by training is also sparse. Therefore, the real time property of the classifier in a dynamic vision task is improved.
Need to check novelty before this filing date? Find Prior Art

Description

technical field

[0001] The invention belongs to the field of machine vision and relates to a classifier design method for classifying dynamic visual information. Background technique

[0002] With the development of pattern recognition and machine learning technology, the application of machine vision in real life is increasing. The main method is to use the camera to obtain dynamic video information, and then use the computer to simulate the human visual function, process and understand the collected visual information. Because machine vision has the characteristics of fast processing speed and large amount of information, it has a wide range of applications in identity authentication, object detection and recognition, robots, and automotive assisted driving systems.

[0003] At present, dynamic vision has made great progress in the field of tracking and recognition. From the perspective of the application of machine vision as a practical application of optomechanical int...

Examples

Embodiment Construction

[0025] Various details involved in the technical solution of the present invention will be described in detail below in conjunction with the accompanying drawings. It should be pointed out that the described embodiments are only intended to facilitate the understanding of the present invention, rather than limiting it in any way.

[0026] The present invention is achieved through the following technical solutions, comprising the following steps:

[0027] Step 1: The video taken by the user in the common environment of the dynamic vision system. The video information must include the target to be recognized and the background environment in normal use.

[0028] Step 2: The user manually collects a small number of samples in the video, including positive samples that recognize the target and negative samples that do not contain the target.

[0029] Step 3: The computer automatically resamples the given video to obtain many samples without category information.

[0030] Step 4:...