{08/09} · Imaging

All work

SE-Conformer for Breast Cancer Detection in Histopathology Images

Role
Second author
Venue
IEEE Access
Status
Published
Year
2025

With Lekha S Nair, Vice Chairperson, Amrita School of Computing

Can local and global image features be fused to detect malignancy in breast histopathology?

A residual convolution block with Squeeze and Excitation attention, joined with Conformer modules to fuse local and global image features.

What the record supports

Claim

Accuracy, precision and recall of about 0.97 for malignancy detection in breast histopathology images.

Evidence

Cited 14 times in Scopus, and used as a comparison baseline in later breast histopathology studies.

Next project · warp onSafety by Design and the Grok Imagine Sexual Deepfake Crisis