Multi-channel topic model-based editable garment image search method

A topic model and image search technology, applied in the field of image processing, can solve problems such as inability to effectively express semantic information and single feature forms, and achieve the effects of overcoming the semantic gap, improving semantic similarity, and optimizing accuracy

Inactive Publication Date: 2018-04-03
SHANGHAI JIAO TONG UNIV
View PDF3 Cites 13 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Aiming at the shortcomings of the single feature form extracted by the prior art and the inability to effectively express semantic information, the present invention proposes an editable clothing image search method based on a multi-channel topic

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
  • Multi-channel topic model-based editable garment image search method
  • Multi-channel topic model-based editable garment image search method
  • Multi-channel topic model-based editable garment image search method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0013] This embodiment includes the following steps:

[0014] Step 1) Image preprocessing: Before extracting the visual descriptor of the clothing sample image, the clothing area should be located, which can reduce the noise brought by the image background and enhance the robustness of the algorithm. Here, the R-CNN detector is trained to identify the clothing main body area and model torso area in the image. Further, use sliding windows of different scales to divide the region into multiple overlapping sub-regions, such as figure 1 shown.

[0015] Step 2) Feature extraction: For each image sub-region, extract visual features such as SIFT features, FilterBank features, and color histogram features, which describe the shape, texture, and color information of clothing, respectively. At the same time, keywords are extracted from the product description of clothing as text features. Further, K-means clustering and local linear coding (LLC) are used to convert the above descript...

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 multi-channel topic model-based editable garment image search method. The method comprises the steps of firstly, finding a main body region of a garment commodity in a picture by using an object detection method, extracting multiple descriptors from the main body region and quantizing the descriptors into vectors in bag-of-word forms by using a bag-of-word model; secondly, according to search conditions, editing and modifying weights of visual words, fusing the vectors in the bag-of-word forms into retrieval features capable of describing high-level semanteme of the garment commodity by using a pre-trained multi-channel topic model, and establishing indexes; and during online detection, calculating vector similarity of a to-be-queried commodity sample image and animage of a database, and taking the commodity with the highest similarity as a search result. The information of vision, commodity text attributes and the like of the to-be-queried commodity can be re-edited; the demand of a user on the commodity is described more accurately; and the user-expected commodity is searched for through the multi-channel topic model.

Description

technical field [0001] The invention relates to a technology in the field of image processing, in particular to an editable clothing image search method based on a multi-channel theme model. Background technique [0002] In recent years, content-based image search (CBIR) technology to retrieve apparel products has become a research hotspot. Because it is difficult to extract the user's retrieval intention by image search mode, the model for extracting features from visual images is too simple to establish image vision. Efficient association of features and semantic information. As a result, users still need to input additional text information to make up for the missing semantics. At present, there have been many works on content-based clothing retrieval, mainly focusing on three issues: 1) how to judge the type and attributes of clothing samples without additional text information; 2) how to extract and fuse the underlying visual information and Clothing features of high-...

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): G06F17/30G06K9/46G06K9/62
CPCG06F16/532G06F16/583G06V10/464G06F18/29
Inventor 周正中张丽清
Owner SHANGHAI JIAO TONG UNIV
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