GBCapsNet: A calibrated capsule network for automated gallbladder disease diagnosis via ultrasound imaging

Madhu Golla, Hareesha Katiganere Siddaramappa, Puvvala Jogeeswara Venkata Naga Sai, Sai Karthik Adla, Chandrika Naga, Pradeep Nijalingappa · 2026

Read GBCapsNet: A calibrated capsule network for automated gallbladder disease diagnosis via ultrasound imaging by Madhu Golla, Hareesha Katiganere Siddaramappa, Puvvala Jogeeswara Venkata Naga Sai, Sai Karthik Adla, Chandrika Naga, Pradeep Nijalingappa (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0348705.txt.

Description

Madhu Golla and a large team built a capsule network, a vision architecture that tracks parts and wholes, to diagnose gallbladder disease from ultrasound images automatically. Reading these scans by hand is slow and varies between operators. Their tailored model, GBCapsNet, adds a custom routing method for sharper class separation and a calibration step that aligns the model's confidence with reality.

Scores under favourable splitting reached near perfection, but the authors flag the catch themselves: images from the same patient could sit on both sides of the split, inflating results. True patient-level separation and outside datasets are still needed before any clinical claim. Calibration did improve the honesty of the model's probabilities. The work stands as an early, carefully hedged exploration of capsule networks for this organ.

Catalog record

CREATORMadhu Golla, Hareesha Katiganere Siddaramappa, Puvvala Jogeeswara Venkata Naga Sai, Sai Karthik Adla, Chandrika Naga, Pradeep Nijalingappa
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE41583
LICENSECC_BY
RIGHTSCC BY
SOURCEPLOS

Source: https://doi.org/10.1371/journal.pone.0348705

Browse the full library · About the Tower of Babel