← All updates
Publication
Lesion localisation in digital breast tomosynthesis with deformable transformers using 2.5D information
Presented at SPIE Medical Imaging 2024 (Computer-Aided Diagnosis), San Diego.
Work by Zhikai Yang, Tianyu Fan, Örjan Smedby, and Rodrigo Moreno.
Abstract
The study adapts a transformer-based method to localise lesions in digital breast tomosynthesis (DBT) images, with the transformer-based formulation removing the need for non-maximum-suppression post-processing.
Integrated deformable-convolution detection transformers can better capture small-size lesions, and transfer learning is added to tackle the issue of the lack of annotated data from DBT. The experimental results demonstrate that the proposed method performs better than all comparison methods.
