Progression-guided spatiotemporal memory transformers for accurate and consistent longitudinal brain tumor segmentation
Sandeep Kumar Mathivanan, Shamala K Subramaniam, Dafik, Sunder R, Sangeetha SKB, Siva Shankar S · 2026
Read Progression-guided spatiotemporal memory transformers for accurate and consistent longitudinal brain tumor segmentation by Sandeep Kumar Mathivanan, Shamala K Subramaniam, Dafik, Sunder R, Sangeetha SKB, Siva Shankar S (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/pmc-13434621.txt.
Description
Sandeep Kumar Mathivanan's team proposed a 2026 model for tracing brain tumors across repeat MRI scans, where most tools treat each scan alone. Treating scans one by one can make a tumor look jagged or jumpy over time, so their design, PGSMT, carries a memory of the tumor from one scan to the next. It aligns shape across time, allows real change, and sharpens edges with a transformer encoder that reads spatial links.
Tested on a longitudinal brain tumor benchmark, the model beat plain neural, mixed, and transformer rivals on enhancing tumor, core, and whole tumor scores, with less volume swing between scans. Clinicians and imaging teams get a steadier way to follow growth and judge whether treatment is working.
Catalog record
| CREATOR | Sandeep Kumar Mathivanan, Shamala K Subramaniam, Dafik, Sunder R, Sangeetha SKB, Siva Shankar S |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 78 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |