Topic recognition

[prianka-indla]

Title: HipMRI Study on Prostate Cancer using 2D UNet Model (Easy Difficulty)

Description: This pull request adds a U-Net-based deep learning model for prostate MRI segmentation, including the following components:

Model Architecture (modules.py):

• Implements a U-Net with an encoder-decoder structure and skip connections for accurate segmentation.
• Uses dropout for regularisation and combines Binary Cross-Entropy (BCE) with Dice loss for optimisation.

Data Processing (dataset.py):

• ProstateMRIDataset class loads and preprocesses MRI scans and masks.
• Data augmentation techniques like flips and rotations are applied to improve robustness.

Train and Validate Script (train.py):

• Initisialise model, load data and train model to accurately predict image segment
• Output 'Dice Score vs Epoch' plot and 'Train and Validate Loss vs Epoch' plot to identify accuracy and efficiency of model

Prediction Script (predict.py):

• Loads the trained model for test inference and visualises results.
• Displays the original image, predicted segmentation, and ground truth, saving the outputs for easy comparison.

ReadMe:

• Includes instructions for environment setup, running the model, and testing.

[yexincheng]

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

1. Solves problem: Very well-documented and detailed report. The average DICE score is lower than 0.75 and the DICE of individual labels were not reported. In Figure 1 (dice vs epochs), it is unclear which dataset was used, validation or training. Why did the dice decrease as the loss decreased?
2. Implementation: Hyperparameters were hard-coded in train.py. Otherwise all great.
3. Design: Well-designed.
4. Comment: Sufficient and detailed comments. The pull request title should be more informative.
5. Difficulty: Easy

Note:

1. Wrong branch for pull request. Should be topic-recognition rather than main.
2. Model checkpoint, pycache and HPC log files should not be committed.

[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. Unclear which dataset was used, validation or training	-1
	Driver Script present.
	File structure present.
	Good Usage & Demo & Visualisation & Data usage.
	Module present.
	Commenting present.
	No Data leakage found.
	Difficulty : Easy. UNet2D (Easy Difficulty)	-10
Commit Log
	Good Meaningful commit messages.
	Good Progressive commits.
Documentation
	Readme :Good.
	Model/technical explanation :Good.
	Description and Comments :Acceptable. Missing references.	-1
	Markdown used and PDF submitted. PDF checked.
Pull Request
	Pull Request has problems: Late submission.	-2
	Feedback action required: Feedback marks possible +2 if the requested changes are made. Remove cache files for mearge.	-2
	Request Description is adequate.	-1
TOTAL		-18

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

[gayanku]

Feedback marks possible +2 if the requested changes are made (see above).

[shakes76]

Approved extension +2