A Two-Stage Lightweight Deep Learning Framework for Mass Detection and Segmentation in Mammograms Using YOLOv5 and Depthwise SegNet

Deleted Journal, Journal Year: 2025, Volume and Issue: unknown

Published: March 14, 2025

Abstract Ensuring strict medical data privacy standards while delivering efficient and accurate breast cancer segmentation is a critical challenge. This paper addresses this challenge by proposing lightweight solution capable of running directly in the user’s browser, ensuring that never leave computer. Our proposed consists two-stage model: pre-trained nano YoloV5 variation handles task mass detection, neural network model just 20k parameters an inference time 21 ms per image problem. highly terms speed memory consumption was created combining well-known techniques, such as SegNet architecture depthwise separable convolutions. The detection manages mAP@50 equal to 50.3% on CBIS-DDSM dataset 68.2% INbreast dataset. Despite its size, our produces high-performance levels (81.0% IoU, 89.4% Dice) (77.3% 87.0%

Language: Английский

A modified MultiResUNet model with attention focus for breast cancer detection

Computers & Electrical Engineering, Journal Year: 2025, Volume and Issue: 124, P. 110416 - 110416

Published: May 1, 2025

Language: Английский