---
tags: [medical-imaging, brain-tumor, wavelet, interpretability]
---

# LDW-CNet: Learnable Discrete Wavelet CNN for Brain MRI Classification

## Results

| Model | Accuracy | Macro F1 | Macro AUC |
|-------|----------|----------|-----------|
| PlainCNN | 59.14% | 0.5265 | 0.8142 |
| FixedDWT_CNN | 67.77% | 0.6312 | 0.8481 |
| LearnDWT_NoLwc | 60.15% | 0.5565 | 0.8356 |
| LearnDWT_SE_NoLwc | 64.97% | 0.5924 | 0.8392 |
| LDWCNet_Full | 68.02% | 0.6324 | 0.8674 |

## Architecture
Learnable DWT → SE Block → Residual CNN → Classifier

## Paper
LDW-CNet: A Learnable Discrete Wavelet Transform based Residual CNN