Screening method, device, server and storage medium for blind spot detection training set

A screening method and training set technology, applied in neural learning methods, instruments, biological neural network models, etc., can solve the problems of time-consuming, labor-intensive, non-uniform training set standards, improve recognition accuracy, reduce costs, and optimize model parameters. Effect

Active Publication Date: 2021-12-31
TIANJIN SOTEREA AUTOMOTIVE TECH LMITED CO
View PDF13 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In view of this, the present invention aims to propose a screening method, device, server and storage medium for a blind spot detection training set, so as to solve the problem that the training set of a convolutional neural network for blind spot detection in the prior art can only be manually marked. Time-consuming and labor-intensive technical issues with inconsistent standards

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
  • Screening method, device, server and storage medium for blind spot detection training set
  • Screening method, device, server and storage medium for blind spot detection training set
  • Screening method, device, server and storage medium for blind spot detection training set

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0034] figure 1 A flow diagram of a screening method of a blind zone detection training set provided by the embodiment of the present invention, and a screening method of a blind zone detection training set provided by an embodiment of the present invention is suitable for screening of image training sets for blind zone detection, especially It is a case where the image training set of the blind zone detects can be screened, and the screening method of the blind zone detection training set can be performed by the screening device of the blind zone detection training set, see figure 1 The screening method of the detection training set of the blind zone, including:

[0035] S110 inputs images to semantic segmentation neural networks.

[0036] The image acquisition device is typically acquired by the image acquisition device mounted on a particular position on the vehicle, and the acquired image is discriminated by convolutional neural network, identifying a target object that may a...

Embodiment 2

[0049] Figure 4 The flow schematic of the screening method of the blind zone detection training set provided by the second embodiment of the present invention, the present embodiment is optimized by the above embodiment, specifically, and specifically maps the output matrix to the base image template, further optimization is: calculation The relative feature ratio of each element in the output matrix in the output matrix generates a relative matrix according to the relative feature ratio; a feature image is generated according to the relative matrix and the output image.

[0050] See Figure 4 The screening method of the detection training set of the blind zone, including:

[0051] S210, input images to a semantic segmentation neural network.

[0052] S220, acquire the semantic segmentation neural network of features identify the output matrix of the anti-roll layering, the output matrix for characterizing the feature strength corresponding to each pixel of the output image.

[00...

Embodiment 3

[0064] Image 6 The flow diagram of the screening method of the blind zone detecting training set provided by the third embodiment of the present invention, the present embodiment is optimized according to the above embodiment, specifically, the method can also increase the following steps: acquiring the semantic segmentation neural network The output result; the image input target detects the neural network; the output result of the neural network according to the semantic division determination of the coordinate position corresponding to the target recognition result is determined according to the coordinate position. See Image 6 The screening method of the detection training set of the blind zone, including:

[0065] S310, input images to a semantic split neural network.

[0066] S320, acquire the semantic segmentation of neural network characteristics to identify the output matrix of the anti-roll layering.

[0067] S330, project the output matrix to the base image template to...

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 provides a screening method, device, server and storage medium for blind spot detection training set, wherein the method includes: inputting an image into a semantic segmentation neural network; obtaining feature recognition deconvolution in the semantic segmentation neural network The output matrix of the layer, the output matrix is ​​used to represent the feature intensity corresponding to each pixel of the output image; the output matrix is ​​mapped to a basic image template to form a feature image, and the basic image template is used to highlight the feature intensity Calculate the mean value of the image feature values ​​of the preset region of interest in the feature image, compare the mean value of the feature values ​​with the preset mean value threshold, and determine when the mean value is not greater than the preset mean value threshold. The above images are added to the blind spot detection training set. The unified, objective and intuitive prediction results of the data set can be obtained, and the targets with poor feature extraction and uncertainties can be added to the next training.

Description

Technical field [0001] The present invention belongs to the field of computer image recognition, in particular, to a screening method, apparatus, server, and storage medium of a blind zone detection training set. Background technique [0002] Car blind zone detection is a car intelligent security technology that can intelligently identify pedestrians and other vehicles in the blind zone through intelligent monitoring cameras and other auxiliary equipment (radar, alarms, display, etc.) installed on the vehicle. According to this, it will be prompted to eliminate the line of viewing, and improve driving safety. [0003] Currently, automotive blind zone detection typically utilizes the camera to take the corresponding image, and use the convolutional neural network to identify the various targets that appear in the image. [0004] In the process of implementing the present invention, the inventors have discovered the following technical problems: When using the convolutional neural ...

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): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/214
Inventor 徐显杰李涛
Owner TIANJIN SOTEREA AUTOMOTIVE TECH LMITED CO
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap