EE Seminar: Automatic breast lesion classification by joint neural analysis of mammography and ultrasound
https://us04web.zoom.us/j/9811616388?pwd=UGlnWVFFT3lFVkFpZlNmVHFNUXRVdz09
Passcode: 5F4CPq
Speaker: Gavriel Habib
M.Sc. student under the supervision of Prof. Nahum Kiryati and Dr. Arnaldo Mayer
Tuesday, November 24th, 2020, at 15:00
Automatic breast lesion classification by joint neural analysis of mammography and ultrasound
Abstract:
Breast cancer is the most common cancer in women worldwide. Mammography is a frequent diagnostic approach with proven mortality reduction and early disease treatment benefits. However, as it suffers from poor lesion visibility in dense breasts, radiologists are using breast ultrasound as a complementary imaging modality. Yet, existing computer-aided diagnosis (CAD) systems for the breast are generally based on a single modality.
In this research, we propose a deep learning-based method for classifying breast cancer lesions from their respective mammography and ultrasound images. The proposed approach is based on a GoogleNet architecture, fine-tuned for our own dataset in two training steps. First, a distinct neural network is trained separately for each modality, generating high-level features. Then, the aggregated features originating from each modality are used to train a multimodal network to provide the final classification. Our approach outperforms state-of-the-art mono-modal models and performs similarly to an average radiologist, surpassing two out of four radiologists participating in a reader study. These results suggest that our model may become a valuable decision support tool for radiologists.
השתתפות בסמינר תיתן קרדיט שמיעה = עפ"י רישום שם מלא + מספר ת.ז. בצ'אט
