Object characteristic processing method and device, storage medium and electronic equipment

A technology of object features and feature maps, applied in character and pattern recognition, instruments, computer components, etc., can solve problems that affect accuracy

Active Publication Date: 2018-06-29
BEIJING SENSETIME TECH DEV CO LTD
View PDF1 Cites 21 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Since the global feature represents the high-level semantic features of the image, detailed information in the image that is very useful for completing

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 characteristic processing method and device, storage medium and electronic equipment
  • Object characteristic processing method and device, storage medium and electronic equipment
  • Object characteristic processing method and device, storage medium and electronic equipment

Examples

Experimental program
Comparison scheme
Effect test

Example Embodiment

[0040] Example one

[0041] figure 1 It is a flowchart showing an object feature extraction method according to the first embodiment of the present invention.

[0042] Reference figure 1 In step S110, feature data of multiple scales of the target object is obtained from the image to be inspected.

[0043] Here, the image to be inspected may be a static image or a video frame image containing the target object. The target object may be an object with a visible shape such as pedestrians, vehicles, animals, and flying objects.

[0044] The feature data of multiple scales of the target object can be extracted from the image to be inspected by an applicable image processing method. Here, multiple scales can be understood as multiple scale resolutions. The feature data of any scale can be, for example, the texture feature data of the image, the color feature data of the image, the shape feature data of the object, or, from another perspective, the The feature vector matrix of the image ex...

Example Embodiment

[0051] Example two

[0052] figure 2 It is a flowchart showing an object feature extraction method according to the second embodiment of the present invention.

[0053] For ease of description, in the present disclosure, it is assumed that the aforementioned multiple scales are N scales, and N is an integer greater than 1.

[0054] In addition, according to this embodiment, the feature data of any scale includes the first feature map corresponding to the multiple feature channels of the scale. Here, the multiple feature channels may correspond to predetermined multiple image features or object features to represent the degree of response of the image to the image features or object features corresponding to each feature channel.

[0055] Reference figure 2 In step S210, a first feature map of N scales of the target object is obtained from the image to be inspected.

[0056] For example, the image to be inspected may be subjected to multiple convolutions and multiple down-sampling poo...

Example Embodiment

[0079] Example three

[0080] image 3 It is a flowchart showing an object feature extraction method according to the third embodiment of the present invention.

[0081] In the object feature extraction method of the third embodiment of the present invention, the first neural network for generating fusion feature data and / or the second neural network for feature extraction can be used to perform the processing of the corresponding steps.

[0082] Reference image 3 In step S310, the feature data of multiple scales of the target object is obtained from the image to be inspected through the second neural network.

[0083] Specifically, the pre-trained second neural network may be used to obtain feature data of multiple scales of the target object from the image to be inspected. The second neural network can obtain multiple scales of feature data of the target object by performing multiple convolutions and pooling on the image to be inspected.

[0084] As mentioned above, optionally, the ...

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 embodiment of the invention provides an object characteristic processing method, an object characteristic processing device, a storage medium and electronic equipment. An object characteristic extraction method comprises the steps of acquiring characteristic data of multi-scales of a target object from a to-be-detected image; generating respective attention heat data of each scale according tothe characteristic data of each scale, wherein the attention heat data represent attention heat of each attention part region of the target object; and acquiring fusion characteristic data of the target object according to the characteristic data of each scale and the attention heat data of each scale. Therefore, the fusion characteristic data fused with the characteristic data of details and overall semantics and the attention heat data are obtained, thereby being beneficial to subsequently and accurately completing an image processing task based on the fusion characteristic data.

Description

technical field [0001] Embodiments of the present invention relate to artificial intelligence technology, and in particular to an object feature processing method, device, computer-readable storage medium, and electronic equipment. Background technique [0002] For solutions such as object attribute detection tasks, object recognition tasks, etc., global features are usually extracted directly from images, and corresponding tasks are completed according to the extracted global features. Since the global feature represents the high-level semantic features of the image, detailed information in the image that is very useful for completing the aforementioned tasks may be lost, such as clothing stripes, whether to wear glasses, etc., thus affecting the accuracy of the aforementioned tasks. Contents of the invention [0003] The purpose of the embodiments of the present invention is to provide an object feature extraction technology. [0004] According to the first aspect of an...

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): G06K9/62G06K9/46
CPCG06V10/464G06F18/213G06F18/253
Inventor 赵海宇刘希慧邵静伊帅闫俊杰王晓刚
Owner BEIJING SENSETIME TECH DEV CO LTD
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