Abstract:
Automated breast ultrasound (ABUS) is a promising breast screening method. Unlike B-mode 2D ultrasound, ABUS provides 3D breast views and operator-independent image acquisition.
ABUS image review is time-consuming and mistakes can happen. We use an innovative 3D convolutional network for ABUS automated cancer detection to speed up reviewing and achieve high detection sensitivity with low false positives (FPs).
We use densely deep supervision and multi-layer features to increase detection sensitivity. We suggest a threshold loss to present a voxel-level adaptive threshold for cancer vs. non-cancer with high sensitivity and low false positives.
Our network is validated by a dataset of 219 patients with 614 ABUS volumes, including 745 cancer regions, and 144 healthy women with 900 volumes without abnormal findings.
Extensive experiments show our method has 95% sensitivity with 0.84 FP per volume. The proposed ABUS-based breast cancer detection network has high sensitivity and low false positives.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
