Robust and Real-Time Bangladeshi Currency Recognition: A Dual-Stream MobileNet and EfficientNet Approach
Subreena 1 , Mohammad Amzad Hossain 1 , Mirza Raquib 1,2 , Saydul Akbar Murad 3 , Farida Siddiqi Prity 4 , Muhammad Hanif 1 , Nick Rahimi 3,* · 2026
Read Robust and Real-Time Bangladeshi Currency Recognition: A Dual-Stream MobileNet and EfficientNet Approach by Subreena 1 , Mohammad Amzad Hossain 1 , Mirza Raquib 1,2 , Saydul Akbar Murad 3 , Farida Siddiqi Prity 4 , Muhammad Hanif 1 , Nick Rahimi 3,* (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/arxiv-2602-07015.txt.
Description
Subreena and six co-authors build a banknote reader for Bangladeshi notes, aimed above all at blind users. Blind users now depend on others to name notes, which puts them at risk of fraud. The team first built a new banknote data set with both controlled and real-world scenes. Then they added four more data sets to widen coverage.
Their hybrid model joins two light convolutional nets for feature extraction. A multilayer perceptron classifier then scores the notes, keeping compute cost low for small devices. Scores hit about 98 percent on controlled data, about 93 percent on complex backgrounds, and about 95 percent combined. Five-fold cross-validation across seven metrics backs those scores. Explainable AI tools like LIME and SHAP add transparency, supporting trust in assistive use.
Catalog record
| CREATOR | Subreena 1 , Mohammad Amzad Hossain 1 , Mirza Raquib 1,2 , Saydul Akbar Murad 3 , Farida Siddiqi Prity 4 , Muhammad Hanif 1 , Nick Rahimi 3,* |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 68 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-07015