Garment Style Recognition Method Based on Contour Curvature Feature Points and Support Vector Machine

Abstract

The invention relates to a clothing style recognition method based on contour curvature feature points and a support vector machine. Firstly, the clothing outline is obtained by preprocessing the clothing image, and then the contour curvature feature points of the clothing outline are extracted as feature vectors, and finally based on the support vector Automatic clothing style recognition. The preprocessing of clothing images includes clothing image segmentation, edge detection, tilt correction, and contour curve smoothing; extracting contour curvature feature points as feature vectors refers to extracting point sets with larger curvature from the clothing contour curve to represent the main points of the contour shape. The features are arranged in a certain order to form a feature vector; the recognition of clothing styles based on support vector machines refers to the classification of clothing styles by using the support vector machine classification method. The clothing style recognition method proposed by the invention can make the clothing style recognition reach an accuracy rate of more than 86%, and the recognition time of each sample is relatively short, and has the characteristics of being fast and accurate.

Description

technical field [0001] The invention belongs to the field of clothing style recognition, and relates to a clothing style recognition method based on contour curvature feature points and support vector machines, in particular to a clothing contour image obtained after image preprocessing, and based on contour curvature feature points and support vectors Machine-based clothing style recognition method. Background technique [0002] With the growing maturity of e-commerce, online shopping for clothing has become increasingly common. However, how to retrieve favorite clothing styles from the massive clothing images on the online shopping platform has always puzzled operators and consumers. The retrieval of clothing styles on the current online shopping platform is based on text retrieval. On the one hand, it takes a lot of manual labor to make labels; clothing. The computer recognition technology of clothing style is expected to solve the above problems. Consumers provide sa...

AI Technical Summary

Problems solved by technology

The shape feature description operators used in these studies (Fourier descriptor, wavelet Fourier descriptor and Hu invariant moments, etc.) can describe the shape features of clothing outlines, but they cannot be intuitively related to the shape features of clothing outlines. Correspond

Images