Research on Image Retrieval Method of Medical Equipment Based On Multi Feature Hierarchical Fusion

2018 3rd International Conference on Insulating Materials, Material Application and Electrical Engineering

材料科学;无线电电子学;电工学

Du, Debiao^1 ; Yang, Zhixiong^2 ; Xia, Fei^3 ; Wang, Yun^4

Yunnan Provincial Hospital of Traditional Chinese Medicine, Kunming

650021, China^1

Kunming Institute of Physics, Kunming

650223, China^2

State Grid Liaoyang Electric Power Supply Company, Liaoyang

111000, China^3

Beijing Institute of Mechanical Equipment, Beijing

100854, China^4

关键词: Endoscopic image;    Feature retrieval;    Fourier descriptors;    Hierarchical fusions;    Retrieval methods;    Similarity measure;    Texture features;    Weighted fusion;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/452/4/042086/pdf
DOI  :  10.1088/1757-899X/452/4/042086

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

To solve the problem of medical equipment image retrieval, an image retrieval method based on multi-feature hierarchical fusion is proposed. This method uses color moments to calculate the color features of the image, LBP operator to calculate the texture features of the image, Fourier descriptor to calculate the shape features of the image, through hierarchical weighted fusion method to fuse the three types of features within a similarity measure, and finally form the basis for retrieval and judgment. Based on the experimental results of endoscopic images and CT images, the results confirm the effectiveness of the proposed method. With the increasing requirement of recall rate, the precision of the proposed method is obviously higher than that of the three single feature retrieval methods, which is suitable for medical image retrieval.