Project 3: Improved 2D UNet for HipMRI Prostate Segmentation (Normal Difficulty) - s4884308

[prabhjotsingh1313]

Project Summary

Implementation of Improved 2D U-Net for prostate cancer segmentation on HipMRI dataset using PyTorch.

Key Achievements

• Prostate Dice coefficient: 0.9373 (exceeds 0.75 requirement by 24.9%)
• Mean Dice across all classes: 0.9158
• Complete modular implementation with comprehensive documentation

Problem Solved

Project 3: Segment HipMRI Study on Prostate Cancer using 2D Improved U-Net with minimum Dice similarity coefficient of 0.75 on prostate label.

Files Included

• modules.py: U-Net architecture components
• dataset.py: NIfTI data loading and preprocessing
• train.py: Training, validation, and testing pipeline
• predict.py: Inference and visualization
• README.md: Complete documentation
• requirements.txt: Dependencies
• images/: Visualizations (training curves, predictions, overlays)

Results on Test Set

Channel	Dice Coefficient
0 (Background)	0.9952
1	0.9768
2	0.9023
3 (Prostate)	0.9373
4	0.8717
5	0.8113

Mean Dice: 0.9158

Testing Instructions

pip install -r requirements.txt
python train.py --data_path /path/to/HipMRI_2D --epochs 20
python predict.py --data_path /path/to/HipMRI_2D --checkpoint checkpoints/best_model.pth

Student: Prabhjot Singh
Student ID: 48843085
Difficulty: Normal
Requirement Met: Prostate Dice ≥ 0.75 Achieved: 0.9373

[wangzhaomxy]

<This is an initial inspection, no action is required at this point.>

File Organizing: Well-organized files.

Problem Solving:

• The algorithm solves the problem appropriately.
• Accuracy in testing dataset: 0.94.

Model and functions:

• It correctly uses PyTorch to construct the improved UNet 2D models and functions.
• NO data augmentation.
• Properly split and use the train/validation/test datasets.

Code design: Good.

Code comment and docstring:

• Good code comments
• Good function docstrings
• Good header block

Difficulty: Normal.

Additional Comments:

• Good commits
• Good ReadMe design and report content. However, discussion and conclusion sections are expected to be in the report.
• The training and validation curves reflect only two epochs of training, whereas the report claims a total of 20 epochs. There is insufficient evidence to support this statement.
• A stable validation loss of approximately 0.3 is observed over two epochs, corresponding to a Dice score of around 0.7. It is therefore unclear how the testing Dice could reach as high as 0.91. There is insufficient evidence to support these reported results.

[gayanku]

Marking

Good/OK/Fair Practice (Design/Commenting, TF/Torch Usage)
	Good design and implementation.
	Spacing and comments.
	No Header blocks.	-1
Recognition Problem
	OK solution to problem.	-1
	Driver Script present.
	File structure present.
	Good Usage & Demo & Visualisation & Data usage.
	Module present.
	Commenting present.
	No Data leakage found.
	Difficulty : Normal. Normal. ImprovedUnet2D	-5
Commit Log
	Good Meaningful commit messages.
	Good Progressive commits.
Documentation
	Readme :Acceptable.	-2
	Model/technical explanation :Good.
	Description and Comments :Good.
	Markdown used and PDF submitted.
Pull Request
	Successful Pull Request (Working Algorithm Delivered on Time in Correct Branch).
	No Feedback required.
	Request Description is good.
TOTAL		-9

Marked as per the due date and changes after which aren't necessarily allowed to contribute to grade for fairness.
Subject to approval from Shakes