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

Object Classification Method Based on Depth Restoration Information

A technology for depth restoration and object classification, applied in neural learning methods, image analysis, biological neural network models, etc., can solve problems such as lack of geometric information, and achieve improved performance and good performance

Active Publication Date: 2022-02-15
TIANJIN UNIV
View PDF11 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

For humans, it is not difficult to infer the underlying 3D structure from a single image. For computer vision algorithms, it is a very challenging task because there are no specific and reliable features such as geometric information that can be directly utilized.

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 Classification Method Based on Depth Restoration Information
  • Object Classification Method Based on Depth Restoration Information
  • Object Classification Method Based on Depth Restoration Information

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0038] The problem to be solved by the present invention is to design a model for the two-dimensional image classification problem so that it can introduce depth information to improve the classification performance, and only need RGB images instead of real depth images collected by sensors as input during testing .

[0039] The technical scheme that the present invention takes is to adopt the deep learning method based on convolutional neural network, and main steps are as follows:

[0040] (1) Preprocess the dataset. Different data sets use different depth sensors to collect depth information and save them in different formats. The matrix of depth information stored in the database is uniformly converted into a general depth image format, which is used for network training and visualization of generated depth maps. And the color image and the corresponding depth image are composed of an image pair, and a 10-fold cross-validation is constructed for network training and resul...

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 present invention relates to the technical field of object classification and monocular image depth estimation in the field of computer vision. It aims to propose a model that can introduce depth information to improve classification performance, and only needs RGB images instead of real images collected by sensors during testing. Depth image is taken as input, and the present invention, the object classification method based on depth restoration information, steps are as follows: (1) data set is preprocessed; (2) depth restoration model in the construction model; (3) training receives RGB and depth respectively Two image classification models with images as input; (4) Construct the final fusion model and perform training and testing; (5) Migrate the fusion network trained in step 4 to the natural image classification data set; (6) Compare the models in Performance and visualization of image classification on two publicly available datasets. The invention is mainly applied to object classification and monocular image depth estimation in the field of computer vision.

Description

technical field [0001] The invention relates to the technical field of object classification and monocular image depth estimation in the field of computer vision, in particular to a depth estimation method based on a generative confrontation network. Background technique [0002] Image object classification is a fundamental problem in computer vision research, and it is also the basis for other high-level vision tasks such as image segmentation, object tracking, and behavior analysis. Since the color RGB image is a two-dimensional projection of the real three-dimensional world, a flat image may correspond to countless actual scenes in the real world. Therefore, the depth information is inevitably lost. Depth information can reflect geometric information that 2D images do not have, and is of great significance for 3D scene reconstruction, gesture recognition, human body pose estimation, etc. [1] . The 2D information represented by the RGB image and the depth information re...

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): G06T7/50G06T7/90G06N3/08G06N3/04G06V10/80
CPCG06N3/08G06T7/50G06T7/90G06N3/045G06F18/253
Inventor 侯春萍管岱杨阳郎玥章衡光
Owner TIANJIN UNIV