Breast mass detection in digital mammography based on anchor-free architecture. (June 2021)

• Record Type: Journal Article
• Authors: Cao, Haichao; Pu, Shiliang; Tan, Wenming; Tong, Junyan
• Is Part Of: Computer methods and programs in biomedicine. Volume 205(2021)
• Publication Date: 2021-06
• Subjects: Breast mass detection -- Anchor-free architecture -- Image enhancement method -- Data augmentation method -- Training method; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2021.106033
• Languages: English
• ISSNs: 0169-2607

Abstract

Highlights: To the best of our knowledge, this paper is the first to apply FSAF structure to breast mass detection. We propose a new normalization method and image enhancement method, which reduce FPPI by about 3 times. We have proposed a new data augmentation method that increases TPR by 5.2% and decreases FPPI by 5.83 times. We propose a dynamic update training method based on sample complexity, which can increase TPR by 2.6% and reduce FPPI by 1.82 times. Abstract: Background and objective: Accurate detection of breast masses in mammography images is critical to diagnose early breast cancer, which can greatly improve the patients' survival rate. However, it is still a big challenge due to the heterogeneity of breast masses and the complexity of their surrounding environment. Therefore, how to develop a robust breast mass detection framework in clinical practical applications to improve patient survival is a topic that researchers need to continue to explore. Methods: To address these problems, we propose a one-stage object detection architecture, called Breast Mass Detection Network (BMassDNet), based on anchor-free and feature pyramid which makes the detection of breast masses of different sizes well adapted. We introduce a truncation normalization method and combine it with adaptive histogram equalization to enhance the contrast between the breast mass and the surrounding environment. Meanwhile, to solve the overfitting problem caused by small data size, we propose a natural deformation data augmentation method and mend the train data dynamic updating method based on the data complexity to effectively utilize the limited data. Finally, we use transfer learning to assist the training process and to improve the robustness of the model ulteriorly. Results: On the INbreast dataset, each image has an average of 0.495 false positives whilst the recall rate is 0.930; On the DDSM dataset, when each image has 0.599 false positives, the recall rate reaches 0.943. Conclusions: The experimental results on datasets INbreast and DDSM show that the proposed BMassDNet can obtain competitive detection performance over the current top ranked methods.