Lesion Localization in Digital Breast Tomosynthesis with Deformable Transformers Detection

Digital breast tomosynthesis (DBT) is an advanced 3D breast imaging technique for breast disease diagnosis. However, checking the DBT is labor-intensive for radiologists due to its many slices and varying mass sizes. To solve this problem, employing artificial intelligence methods could reduce the radiologist's workload. We proposed a transformer-based method to localize lesions in digital breast tomosynthesis images. The experimental results demonstrated that the proposed method performs better than the comparison methods. This seminar will introduce my first-year preliminary work on this and its future extensions.

Reviewers: Sanna Persson and Eva Breznik