Comparing Different Data Partitioning Strategies for Segmenting Areas Affected by COVID-19 in CT Scans
Anne de Souza Oliveira, Marly Guimarães Fernandes Costa, João Pedro Guimarães Fernandes Costa, Cícero Ferreira Fernandes Costa Filho, Thomas Frauenfelder · 2024
Read Comparing Different Data Partitioning Strategies for Segmenting Areas Affected by COVID-19 in CT Scans by Anne de Souza Oliveira, Marly Guimarães Fernandes Costa, João Pedro Guimarães Fernandes Costa, Cícero Ferreira Fernandes Costa Filho, Thomas Frauenfelder (2024) 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-11674714.txt.
Description
Anne de Souza Oliveira and co-authors test how data splitting changes an AI tool that outlines COVID-19 damage in chest scans, in a 2024 paper. The lab swab stays the gold test, yet scans can back it when swabs read negative in a sick patient. Networks learn to paint the hurt lung zones. How the training scans are split, whole scans or loose slices, can tilt the score.
Two datasets and two network designs were run both ways. Slicing won each time: overlap scores near 85 percent, against about 73 to 74 for whole-scan splits. Hand-drawn lung borders barely beat machine-drawn ones. On a small set of seven scans, the machine's outlines matched the spread between two human reader groups. The authors warn that split choice alone can swell or shrink a tool's grade.
Catalog record
| CREATOR | Anne de Souza Oliveira, Marly Guimarães Fernandes Costa, João Pedro Guimarães Fernandes Costa, Cícero Ferreira Fernandes Costa Filho, Thomas Frauenfelder |
|---|---|
| YEAR | 2024 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 46 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |