Image semantic analysis and retrieval recommendation for clothing based on deep learning
Authors: Xu, H., Bai, M., Wan, T., Xue, T. and Tang, W.
Journal: Fangzhi Gaoxiao Jichukexue Xuebao
Volume: 33
Issue: 3
Pages: 64-72
ISSN: 1006-8341
DOI: 10.13338/j.issn.1006-8341.2020.03.011
Abstract:Aiming at the problems of low segmentation accuracy and poor multi-scale feature fusion of clothing image segmentation algorithm in clothing detection and recommendation method, a deep convolution network clothing feature extraction analysis retrieval and recommendation method based on attention mechanism and multi-scale feature fusion is proposed. Based on the image semantic segmentation algorithm, this method extracts high-level semantic information from clothing images, encodes the extracted features by hash function, and extracts multi-scale features of the image by introducing ASPP, and then merges the extracted features with attention model to build an index database, and finally the function of clothing recommendation with semantic similarity is realized. The experimental results show that this method can effectively improve the segmentation accuracy, solve the problem of multi-scale feature extraction, successfully extract the high-level semantic features in the image, with high segmentation accuracy, and greatly improve the efficiency of semantic recommendation of clothing similarity.
Source: Scopus