Visual Features for Improving Endoscopic Bleeding Detection Using Convolutional Neural Networks
Adam Brzeski, Tomasz Dziubich, Henryk Krawczyk · 2023
Read Visual Features for Improving Endoscopic Bleeding Detection Using Convolutional Neural Networks by Adam Brzeski, Tomasz Dziubich, Henryk Krawczyk (2023) 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/pmc-10748269.txt.
Description
Adam Brzeski, Tomasz Dziubich, and Henryk Krawczyk teach neural networks to spot bleeding in endoscopy video by first showing them what bleeding looks like, in a 2023 paper. Expert knowledge was turned into a set of high-level visual features, each with a detector that paints a grayscale map over the frame wherever the feature fires. Those maps ride alongside the normal color channels into standard image classifiers.
On test data, the feature-enriched networks beat the same networks fed plain color images, for two different architectures. The gain shows domain knowledge still helps deep learning when it is offered as extra eyes rather than rules. The task matters for catching bleeds during live procedures.
Catalog record
| CREATOR | Adam Brzeski, Tomasz Dziubich, Henryk Krawczyk |
|---|---|
| YEAR | 2023 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |