Neural network model based on global and local features for multi-view mammogram classification

Xia, Lili; An, Jianpeng; Ma, Chao; Hou, Hongjun; Hou, Yayi; Cui, Linyang; Jiang, Xuheng; Li, Wanqing; Gao, Zhong-Ke

doi:10.1016/j.neucom.2023.03.028

Cited by 7 publications

(2 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our proposed method not only employs CNNs to extract features from bilateral mammograms but also realizes local-to-global area feature attention correlation learning between the two breasts, overcoming the limitation of the CNN’s local receptive field. Additionally, in terms of model design, current works [ 5 , 8 , 9 , 10 , 11 , 12 , 14 , 33 , 34 , 35 , 36 ] mainly focus on improving the diagnostic accuracy for individual breasts, with only a few methods designing models for four-view analysis [ 4 , 18 , 19 , 37 ]. However, these models fail to fully explore the relationship between bilateral mammograms.…”

Section: Related Workmentioning

confidence: 99%

“…The prevailing approach involves using multi-view image information as the input for models. For works that use ipsilateral views as model inputs, the CC and MLO views are used for feature extraction for cancer diagnosis [ 8 , 9 ], but each view extracts features independently and then combines them, resulting in a lack of information interaction between views, leading to the loss of inter-view information relationships. In 2021, van Tulder et al [ 10 ] introduced an inter-view attention mechanism for information transfer between views.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Breast Cancer Diagnosis Method Based on Cross-Mammogram Four-View Interactive Learning

Wen,

Li,

Yang

2024

Tomography

View full text Add to dashboard Cite

Computer-aided diagnosis systems play a crucial role in the diagnosis and early detection of breast cancer. However, most current methods focus primarily on the dual-view analysis of a single breast, thereby neglecting the potentially valuable information between bilateral mammograms. In this paper, we propose a Four-View Correlation and Contrastive Joint Learning Network (FV-Net) for the classification of bilateral mammogram images. Specifically, FV-Net focuses on extracting and matching features across the four views of bilateral mammograms while maximizing both their similarities and dissimilarities. Through the Cross-Mammogram Dual-Pathway Attention Module, feature matching between bilateral mammogram views is achieved, capturing the consistency and complementary features across mammograms and effectively reducing feature misalignment. In the reconstituted feature maps derived from bilateral mammograms, the Bilateral-Mammogram Contrastive Joint Learning module performs associative contrastive learning on positive and negative sample pairs within each local region. This aims to maximize the correlation between similar local features and enhance the differentiation between dissimilar features across the bilateral mammogram representations. Our experimental results on a test set comprising 20% of the combined Mini-DDSM and Vindr-mamo datasets, as well as on the INbreast dataset, show that our model exhibits superior performance in breast cancer classification compared to competing methods.

show abstract