Abstract

With the increasing scale of e-commerce, the complexity of image content makes commodity image classification face great challenges. Image feature extraction often determines the quality of the final classification results. At present, the image feature extraction part mainly includes the underlying visual feature and the intermediate semantic feature. The intermediate semantics of the image acts as a bridge between the underlying features and the advanced semantics of the image, which can make up for the semantic gap to a certain extent and has strong robustness. As a typical intermediate semantic representation method, the bag-of-visual-words (BoVW) model has received extensive attention in image classification. However, the traditional BoVW model loses the location information of local features, and its local feature descriptors mainly focus on the texture shape information of local regions but lack the expression of color information. Therefore, in this paper, the improved bag-of-visual-words model is presented, which contains three aspects of improvement: (1) multiscale local region extraction; (2) local feature description by speeded up robust features (SURF) and color vector angle histogram (CVAH); and (3) diagonal concentric rectangular pattern. Experimental results show that the three aspects of improvement to the BoVW model are complementary, while compared with the traditional BoVW and the BoVW adopting SURF + SPM, the classification accuracy of the improved BoVW is increased by 3.60% and 2.33%, respectively.

Highlights

  • With the increasing of e-commerce, online shopping has become the main way for the public to buy goods

  • The image feature extraction mainly includes the underlying visual feature and the intermediate semantic feature [4, 5]. e underlying visual features mainly refer to the color, texture, and shape of the image

  • Spatial pyramid matching (SPM) adopts a multiscale method to make its structure present a hierarchical pyramid shape. e SPM model is suitable for most scenes, but it lacks particularity, which is not the best representation in the case of spatial distribution with characteristic law. erefore, we propose diagonal concentric rectangular pattern (DCRP) which is suitable for describing the spatial distribution characteristics of commodity images

Read more

Summary

Introduction

With the increasing of e-commerce, online shopping has become the main way for the public to buy goods. 3. Improved Bag-of-Visual-Words Model e traditional BoVW loses the location information of local features, and its local feature descriptors mainly focus on the texture or shape information of local regions but lack the expression of color information. E final local region feature description vector can be obtained by splicing the 64-dimensional SURF vector with the 180-dimensional CVAH, which is 244-dimensional and can effectively describe the color, shape, and texture information of the local region. The BoVW model assigns a local feature vector to a unique visual word closest to it. DCRP effectively introduces the location information of local features in commodity images, which improves the representation ability of BoVW

Experimental Results and Analysis
Method
Conclusions
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call