Painting Image Retrieval Method Based on Color and Texture Features
Jiangqi Hu, Guihua Cui, Xiukai Ruan, Yishan Jiang · 2021
In the paint industry, querying a certain texture image is usually done by employees visually with their personal experience or with the help of a common image retrieval system, which cannot meet the needs of paint companies to query images accurately. In order to improve the accuracy of retrieval, an image retrieval algorithm is proposed for paint images with a wide variety of colors and complex texture information. For color features, a color autocorrelogram is selected; for texture features, a direction-improved uniform local binary pattern that considers the comparison of gray values between neighboring pixels is proposed to enhance texture directional feature recognition. The color and texture features are fused as feature descriptors to retrieve 216 insulated decorative integrated panel images. The experimental results show that the fused features are more suitable for describing particular paint images and have a higher average finding accuracy than other descriptive feature algorithms.