Bleeding Detection and Classification

### Classification Metrics
| Metric (in probability)| Value |
|------------------------|----------|
| Accuracy | 0.7864 |
| Recall (Bleeding) | 0.9602 |
| Recall (Non Bleeding) | 0.6126|
| Recall (Weighted) | 0.7864|
| F1-Score (Bleeding) | 0.8180 |
| F1-Score (Non Bleeding) | 0.7415 |
| F1-Score (Weighted) | 0.7798 |

Detection Metrics

Training Metrics

Confusion Metrics	Confusion Metrics Normalized

F1 Curve	P Curve	PR Curve	R Curve

Detection and Classification

Validation Dataset

Imagename	img- (43).png	img- (44).png	img- (68).png	img- (164).png	img- (200).png
Images
Imagename	img- (448).png	img- (514).png	img- (628).png	img- (687).png	img- (677).png
Images

Interpretability Plots (Cam Plots of 2nd last layer)

Test Dataset 1

Imagename	A0010.png	A0021.png	A0031.png	A0040.png	A0045.png
Cam Images

Test Dataset 2

Imagename	A0164.png	A0256.png	A0300.png	A0384.png	A0473.png
CAM Images

Detection and Classification

Testing Dataset 1

Imagename	A0047.png	A0043.png	A0048.png	A0040.png	A0035.png
Images

Testing Dataset 2

Imagename	A0258.png	A0260.png	A0500.png	A0261.png	A0498.png
Images

Instructions

• Clone the repo
• Install requirements.txt via pip install -r requirements.txt
• Download and extract the model weights from here
• Do necessary changes in test.py file.
• Running the test.py generates the predictions.csv file