End-to-end human parsing and detection optimized for resource-constrained devices
Md Imran Hosen, Tarkan Aydin, Md Baharul Islam · 2025
Read End-to-end human parsing and detection optimized for resource-constrained devices by Md Imran Hosen, Tarkan Aydin, Md Baharul Islam (2025) 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-12783623.txt.
Description
Md Imran Hosen and co-authors design a vision system that parses people in images on weak hardware. Human parsing tags each pixel as a body part or clothing item and ties parts to the right person. Heavy models need helper inputs and big chips, which shuts out cheap devices.
Their end-to-end design folds detection and parsing into one pass. A transformer attention block deepens scene context without a heavy frame. New polygon labels around each person let the model learn detection and parsing at once. The system runs in real time on a Raspberry Pi board and keeps fine detail. Speed without a sharp accuracy loss is the headline gain. Such lean parsing could serve home robots, store cameras, and aid tools in the field.
Catalog record
| CREATOR | Md Imran Hosen, Tarkan Aydin, Md Baharul Islam |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 92 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |