Analyzing Parameter-Efficient Convolutional Neural Network Architectures for Visual Classification
Nazmul Shahadat, Anthony S Maida, Gaochang Wu, Zizhu Fan, Dong Pan · 2025
Read Analyzing Parameter-Efficient Convolutional Neural Network Architectures for Visual Classification by Nazmul Shahadat, Anthony S Maida, Gaochang Wu, Zizhu Fan, Dong Pan (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-12736689.txt.
Description
Nazmul Shahadat and co-authors review, in 2025 work, ways to make image-sorting networks leaner without losing skill. Deep, wide networks sort images well but cost much in time and memory. The review walks through three routes to trim them: rich number systems that share weights, side-wise focus steps, and split filters that break one step into two cheap ones.
The authors set their own past models in that frame. One cut delay and size by using rich layers all through the net. An other split rich filters into steps in a row, which halved the count. Their lean one-way networks matched famed light models on sorting and image sharpening with far fewer parts. The review closes with paths for leaner vision tools.
Catalog record
| CREATOR | Nazmul Shahadat, Anthony S Maida, Gaochang Wu, Zizhu Fan, Dong Pan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 80 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |