Method of Ethnic Minority Costume Image Retrieval Based on Deep Hashing

To solve the problems of low accuracy, high cost, and poor efficiency in image retrieval of ethnic minority costumes, a deep hash image retrieval method combining image generation and image enhancement technologies is proposed. This method utilizes image generation technology to effectively improve the quality of the dataset. It employs corresponding image enhancement techniques for images within different brightness mean ranges to improve the image feature contrast. The deep hash algorithm is used to calculate the hash values of each image feature. By comparing the Hamming distance, the retrieval of ethnic minority costumes is achieved. Experimental results show that, compared to ordinary hash retrieval methods, this method not only improves the success rate of the top 5 image retrievals but also enhances the accuracy and recall rate of image retrieval. It can effectively realize the retrieval and identification of costumes of the Bai, Manchu, Miao, and Uighur ethnic groups, providing an important theoretical basis for the study of ethnic minority cultures, understanding of folk customs, and promotion of ethnic unity.