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

A Scene Classification Method Based on Layer Features in Image

A scene classification and image technology, which is applied in the direction of instruments, computing, character and pattern recognition, etc., can solve the problems of insufficient image discrimination and low representativeness, and achieve the effect of improving the promotion ability

Active Publication Date: 2018-12-07
BEIJING UNIV OF TECH
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0011] The purpose of the present invention is to provide an image classification method based on middle-level feature learning, which uses middle-level features to better simulate human visual characteristics, so as to solve the problems of insufficient image discrimination and low representativeness in image classification

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
  • A Scene Classification Method Based on Layer Features in Image
  • A Scene Classification Method Based on Layer Features in Image
  • A Scene Classification Method Based on Layer Features in Image

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0068] In order to enable those skilled in the art to better understand and use the present invention, the technical solutions of the present invention will be further described below in conjunction with the accompanying drawings.

[0069] 1. Use the camera installed on the robot platform to collect scene images to obtain training images and test images. Mark the object area on the training image, and scale the image block size of the marked area to an 8×8 pixel area, and calculate the 64-dimensional gradient norm feature of this area. Randomly sample the non-marked area of ​​the image, and the obtained image block is also scaled to an 8×8 pixel area, and its 64-dimensional gradient norm feature is calculated.

[0070] 2. Use the gradient norm feature obtained in the marked area as the positive sample of the object template training, and the gradient norm feature obtained in the non-labeled area as the negative sample of the object template training, and use the linear support...

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 a scene classification method based on image middle layer features, which belongs to the field of scene image classification. Firstly, by using a method to quickly find the object area in the image, the image block of the object area is obtained, and the obtained area image block is used as the middle-level feature of the image. Then cluster the middle-level feature image blocks, and use SVM to train a shape template for each cluster, and select n shape templates with the highest rating scores according to the rating index. Finally, scene image classification is achieved based on the matching scores of the test image to n shape templates. The invention solves the problems of insufficient image discrimination and low representativeness in scene image classification.

Description

technical field [0001] The invention relates to an image classification method, in particular to a scene image classification method based on image mid-layer feature learning, which is used for classification of indoor and outdoor natural scene images. Background technique [0002] At present, as more and more intelligent mobile robots enter people's daily life and play an important role in various fields such as military affairs, shopping malls, hospitals, and homes, people's demand for automatic positioning of intelligent mobile robot systems is becoming more and more urgent. Only when the mobile robot knows its own position and the working space accurately can it carry out autonomous movement safely and effectively and serve human beings. As an important research content of image understanding, scene image classification is an important research problem in the field of machine vision and pattern recognition. The scene image classification applied to mobile robots aims to...

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): G06K9/62
CPCG06F18/24
Inventor 杨金福张强张珊珊陈浩
Owner BEIJING UNIV OF TECH