Performance evaluation of AkidaNet converted to spiking domain for the classification of weeds in cotton fields
Mrudula Jeeva, Santhosh Miriala · 2026
Read Performance evaluation of AkidaNet converted to spiking domain for the classification of weeds in cotton fields by Mrudula Jeeva, Santhosh Miriala (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/PMC13529717.txt.
Description
Mrudula Jeeva and Santhosh Miriala convert a vision network into a spiking one for spotting weeds in cotton fields in a 2026 study. Spiking networks mimic brain pulses and sip power. Six weed classes were the target. A compact network called AkidaNet was first trained the normal way. It was then shrunk by quantization and rebuilt as a spiking model with vendor tools.
Plain quantization dropped accuracy to about eighty-eight percent. Training that anticipates shrinking recovered it to about ninety-one percent. After conversion, the spiking model scored about ninety-four percent on test data. That nearly matches the original network's validation form. The authors say such models fit low-power neuromorphic chips. Those chips could ride on field robots for live weed control. Cotton growers lose yield and lint quality to weeds, so cheap automation matters.
Catalog record
| CREATOR | Mrudula Jeeva, Santhosh Miriala |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 28924 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |