DOI:10.1109/STSIVA66383.2025.11156301
This framework learns optimal Gaussian Optical Bandpass Filters that transform high-dimensional spectral signatures into compact, information-rich representations π
Deep Gaussian Optical Bandpass Filter Design for Fermentation Index Estimation in Cocoa Beans π¬π«
A powerful deep learning framework for learnable Optical Bandpass Filter Design! π― Transform spectral signatures into meaningful insights with AI-optimized Gaussian filters πβ¨
FilterDesign is the β core β of this work!
It learns Gaussian optical filters that reduce spectral dimensionality while preserving the most discriminative features π
π¬ Technical approach:
- π‘ Processes raw spectral data (hundreds to thousands of bands)
- π§ Learns optimal Gaussian filter parameters (ΞΌ, Ο)
- βοΈ Reduces dimensionality to physically feasible spectral bands
- π― Achieves state-of-the-art performance with significantly fewer spectral inputs β¨
python deep_learning.py --mode filter_design --learned-bands 6 --epochs 10
python deep_learning.py --mode baseline --epochs 10
python run.py # Tests 3, 6, 11 filters automatically| Mode | Description |
|---|---|
filter_design |
π Primary method - Learnable Gaussian optical filters |
band_selection |
Learnable binary spectral band selection |
binary_band_selection |
Hard binary spectral band selection |
baseline |
Standard full-spectrum approach (no filtering) |
- π§ͺ FilterDesign: Gaussian optical filters with learnable ΞΌ & Ο
- π€ Multiple Backbones: SpectraNet, CNN, LSTM, Transformer, SpectralFormer
- π Two-Stage Training: Joint optimization + filter parameter freezing
- ποΈ FWHM Constraints: Bandwidth limited to physically feasible range (8.25β41.25 nm)
- π― Spectral Efficiency: Drastically reduces dimensionality with minimal performance loss
- π‘ Interpretability: Highlights relevant spectral regions linked to the task
- π§ Modularity: Compatible with diverse neural architectures
- π Validated: Demonstrated superiority over state-of-the-art band selection methods
If you find the OBF-Design useful in your research, please consider citing:
@INPROCEEDINGS{11156301, author={Diaz-Delgado, Laura C. and Monroy, Brayan and Bacca, Jorge and Arguello, Henry}, booktitle={2025 XXV Symposium of Image, Signal Processing, and Artificial Vision (STSIVA)}, title={Deep Gaussian Optical Bandpass Filter Design for Fermentation Index Estimation in Cocoa Beans}, year={2025}, volume={}, number={}, pages={1-5}, keywords={Optical filters;Filters;Biomedical optical imaging;Optical design;Estimation;Optical computing;Optical imaging;Fermentation;Optical sensors;Optical signal processing;fermentation;cocoa;deep learning;spectral imaging;filter design}, doi={10.1109/STSIVA66383.2025.11156301}}
πΈ Keep learning, keep filtering! πΈ