Data acquisition method and device for automatic foreground extraction

A technology of foreground extraction and data acquisition, which is applied in image data processing, instruments, computing, etc., can solve the problems of data diversity, single background, etc., and achieve the effect of huge value of data augmentation

Pending Publication Date: 2019-05-31
创新奇智(上海)科技有限公司
View PDF5 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] 1. It is necessary to train the detection model of the corresponding category in advance;
[0004] 2. Background information will be added at the same time of automatic labeling, resulting in a relatively single background, which will affect the diversity of data

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
  • Data acquisition method and device for automatic foreground extraction
  • Data acquisition method and device for automatic foreground extraction
  • Data acquisition method and device for automatic foreground extraction

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0035] Embodiments of the present invention are described in detail below, and examples of the embodiments are shown in the drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the figures are exemplary and are intended to explain the present invention and should not be construed as limiting the present invention.

[0036] Embodiments of the present invention provide a data collection method and device for automatic foreground extraction, which can realize target detection, semantic segmentation and data enhancement algorithms in the field of computer vision.

[0037] Such as figure 1 As shown, the data collection method of the automatic foreground extraction of the embodiment of the present invention, comprises the following steps:

[0038] Step S1, perform foreground extraction on the video image to be processed to obtain a for...

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 provides an automatic foreground extraction data acquisition method and device, and the method comprises the steps: carrying out the foreground extraction of a to-be-processed video image, and obtaining a foreground image; storing the foreground images and the corresponding masks in pairs as foreground sku information in a sku database for subsequent training; searching sku information of a to-be-trained list according to a training task, reading required foreground sku information from the sku database, and pulling a sample from the sku database as a negative sample; selecting background information from a basic database, and synthesizing the foreground sku information and the background information; and training the synthesized data by using the negative sample. No manual assistance is needed for foreground extraction, and full automation is achieved; Various complex backgrounds can be fused, and the data augmentation value is huge.

Description

technical field [0001] The invention relates to the technical field of image data processing, in particular to a data collection method and device for automatic foreground extraction. Background technique [0002] Existing dataset collection includes two common tasks, one is the detection task, and the other is the semantic segmentation task. Among them, the most common labeling method is manual labeling, which is time-consuming and laborious, and has a low accuracy rate. The existing solution is to use artificial intelligence-assisted labeling methods, such as labeling methods based on tracking algorithms, but this method has the following defects: [0003] 1. It is necessary to train the detection model of the corresponding category in advance; [0004] 2. Background information will be added during automatic labeling, resulting in a relatively single background, which will affect the diversity of data. Contents of the invention [0005] The aim of the present inventi...

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): G06T7/194G06T5/00
Inventor 张发恩赵江华秦永强
Owner 创新奇智(上海)科技有限公司
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