2D Prostate Segmentation with Improved UNet - Student 47222610 #272

ChiaJouLu · 2025-11-02T21:17:57Z

Project Summary

Student ID: 47222610
Project: #3 - HipMRI 2D Prostate Segmentation with Improved UNet
Difficulty: Normal

Implementation

Implemented Improved UNet architecture for automated prostate segmentation from 2D MRI images.

Key Features:

5-level encoder/decoder with residual connections
Instance Normalization for stable training
Context aggregation module (dilated convolutions: 1, 2, 4, 8)
Deep supervision at 5 decoder levels

Files:

modules.py - Model architecture
dataset.py - Data loading
train.py - Training script
predict.py - Testing script
README.md - Documentation
Result_Images/ - Visualizations

Results

Test Performance:

Class	Dice
Background	0.9881
Peripheral Zone	0.9842
Transition Zone	0.9271
Prostate	0.9552

Requirement: ≥ 0.75
Achieved: 0.9552 (27.4% above requirement)

- Use provided utility functions from assignment Appendix B - Implement load_data_2D with Z-score normalization - Implement to_channels for one-hot encoding - Support large dataset: 11,460 train, 660 val, 540 test images - Add comprehensive documentation and test script - Verified on Rangpur with actual HipMRI data

- Implemented the standard Conv2d -> BatchNorm2d -> ReLU sequence repeated twice. - Ensures dimensions are preserved using padding=1. - This serves as the fundamental building unit for the later Encoder and Decoder paths.

- Integrates MaxPool2d (2x2, stride 2) for downsampling. - After downsampling, use DoubleConv to expand channels and refine features (C -> 2C). - Ensures data shape transformation: (N, C, H, W) -> (N, 2C, H/2, W/2).

- Implements the upsampling to double H/W dimensions of the deep features. - Integrates the skip connection by concatenating the upsampled features with the corresponding encoder features. - Applies DoubleConv after concatenation to fuse high-level semantics and low-level detail, while halving the number of channels.

- Implemented a 1×1 convolution layer to map final feature maps to output classes - No activation applied; handled by loss function - Input shape: (N, C_in, H, W) - Output shape: (N, num_classes, H, W)

…dical image segmentation - Implemented full encoder–decoder structure with skip connections - Includes DoubleConv, Downsampling, Upsampling, and OutConv components - Supports variable input/output channels (e.g. 1→4 for HipMRI prostate segmentation) - Forward pass outputs raw logits

…luation

…tion, and new training loop - Added 'load_data_with_resive()' and 'load_labels_with_resive()' to debug and preprocess input images and labels with resizing support. - Updated Dice metric: replaced 'dice_coefficient' with 'dice_coefficient_per_class()' to compute Dice score separately for each class. - Implemented 'train_one_epoch()' function for training loop structure. - Minor refactoring and cleanup for debugging consistency.

…ice)

…tate segmentation

- simplify runner output and final checkpoint; improve the entire things; keep core training logic unchanged

…tatements to README for clarity

…tch_0.png Correcting classes names

…tch_1.png Correcting classes names

…tch_2.png Correcting classes names

…tch_3.png Correcting classes names

…tch_4.png Correcting classes names

ChiaJouLu · 2025-11-03T05:10:44Z

Important Update: Corrected Class Labels

I have corrected an error in the class descriptions after my initial submission.

Corrections made:

Class 1: ~~"Peripheral Zone"~~ → "Body"
Class 2: ~~"Transition Zone"~~ → "Bone"

Files updated:

README.md
train.py
predict.py
Result_Images/ (all visualizations regenerated)

Note: These are documentation corrections only. Model implementation and results remain unchanged.

Turnitin: Unable to resubmit PDF. Please review the corrected version here on GitHub.

wangzhaomxy · 2025-11-21T19:08:02Z

<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 (Dice): 0.95. Meet the requirements.

Model and functions:

It correctly uses PyTorch to construct the improved UNet 2D models and functions.
NO data augmentation.
Properly 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
Excellent ReadMe design and comprehensive report content. However, as a report, discussion and conclusion section is expected to properly summarize your project.
In fact, the HipMRI dataset contains six labels: label 3 corresponds to the bladder, and label 5 corresponds to the prostate. This does not match the description provided in your report.

gayanku · 2025-11-24T23:18:51Z

Marking

Good/OK/Fair Practice (Design/Commenting, TF/Torch Usage)
	Adequate design and implementation.	-1
	Spacing and comments.
	Header blocks.
Recognition Problem
	OK solution to problem.	-2
	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.	-1
	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

ChiaJouLu added 30 commits October 29, 2025 00:19

Create README.md for Improved UNet

d68bfb3

Delete recognition/UNet_47222610/README.md

aabf517

Create README.md for GNN

0a9a6eb

Delete recognition/gnn_47222610/README.md

fe062e6

Create README.md for 2D Improved UNet

d4c3310

Delete recognition/2D_Improved_UNet_47222610/README.md

8683ad9

Add requirements.txt

3250bce

Create project structure with TODO templates

ff9c82a

Update README.md, delete extra words

f10c6ef

Update Modules.py: Implement the basic DoubleConv block for UNet

3783c5b

- Implemented the standard Conv2d -> BatchNorm2d -> ReLU sequence repeated twice. - Ensures dimensions are preserved using padding=1. - This serves as the fundamental building unit for the later Encoder and Decoder paths.

Update Modules.py: Implement the Downsampling block

7da98d5

- Integrates MaxPool2d (2x2, stride 2) for downsampling. - After downsampling, use DoubleConv to expand channels and refine features (C -> 2C). - Ensures data shape transformation: (N, C, H, W) -> (N, 2C, H/2, W/2).

Update Modules.py: Implement the Output Convolution Layer

d70706c

- Implemented a 1×1 convolution layer to map final feature maps to output classes - No activation applied; handled by loss function - Input shape: (N, C_in, H, W) - Output shape: (N, num_classes, H, W)

Update Modules.py: Fix variable name bug in UNet forward pass

7419aaf

Update dataset.py: Add AI statement

9d371fc

Update modules.py: Add AI statement

d3c33dc

Update train.py: add dice_coefficient() function for segmentation eva…

9a49fad

…luation

Update train.py: add validate() evaluation loop (no grad, per-class D…

f351661

…ice)

Update predict.py: add testing & visualization pipeline for UNet pros…

96f64c0

…tate segmentation

Refactor train.py and predict.py

c0f6cd4

- simplify runner output and final checkpoint; improve the entire things; keep core training logic unchanged

Delete recognition/UNet_Prostate_47222610/utils.py

6f00b72

Delete recognition/UNet_Prostate_47222610/requirements.txt

1db6c2e

Cleanup dataset.py: remove unused main() block from dataset.py

352d7d9

Update modules.py: remove unused main() block and relocate AI usage s…

3e766d8

…tatements to README for clarity

Update predict.py: update docstrings for consistency across functions

ea8fea8

Update train.py: update docstrings for consistency across functions

559ea50

Update README.md: complete Problem Description section

af82ae4

ChiaJouLu added 20 commits November 3, 2025 06:33

Delete recognition/UNet_Prostate_47222610/prediction_batch_0.png

225f9b4

Delete recognition/UNet_Prostate_47222610/prediction_batch_1.png

5a474da

Delete recognition/UNet_Prostate_47222610/prediction_batch_2.png

1dc8428

Delete recognition/UNet_Prostate_47222610/prediction_batch_3.png

c707c01

Delete recognition/UNet_Prostate_47222610/prediction_batch_4.png

8999cef

Delete recognition/UNet_Prostate_47222610/training_curves.png

93b5586

Create .gitkeep for later images upload

e9865f1

Move images to Result_Images/ folder

b7f8188

Delete recognition/UNet_Prostate_47222610/Result_Images/.gitkeep

d4bb82a

Update README.md: add visualization section and complete everything

d005bc8

Update README.md: correcting classes names

dba11e1

Update train.py: correcting classes names

93bc7da

Update predict.py: correcting classes names

b13b121

Update predict.py: correcting labels names

a415f8e

Delete recognition/UNet_Prostate_47222610/Result_Images/prediction_ba…

7fd1cef

…tch_0.png Correcting classes names

Delete recognition/UNet_Prostate_47222610/Result_Images/prediction_ba…

61244c8

…tch_1.png Correcting classes names

Delete recognition/UNet_Prostate_47222610/Result_Images/prediction_ba…

6c25c2d

…tch_2.png Correcting classes names

Delete recognition/UNet_Prostate_47222610/Result_Images/prediction_ba…

45e2534

…tch_3.png Correcting classes names

Delete recognition/UNet_Prostate_47222610/Result_Images/prediction_ba…

d17278e

…tch_4.png Correcting classes names

Add correcting result images

6e53d98

hanemma7moud added _Easy _ImprovedUnet2D _Normal and removed _Easy labels Nov 10, 2025

wangzhaomxy added checked-1st Uploaded_PDF labels Nov 21, 2025

gayanku added the SECOND_MARK label Nov 24, 2025

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

2D Prostate Segmentation with Improved UNet - Student 47222610 #272

2D Prostate Segmentation with Improved UNet - Student 47222610 #272

Uh oh!

ChiaJouLu commented Nov 2, 2025

Uh oh!

ChiaJouLu commented Nov 3, 2025

Uh oh!

wangzhaomxy commented Nov 21, 2025

Uh oh!

gayanku commented Nov 24, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

2D Prostate Segmentation with Improved UNet - Student 47222610 #272

Are you sure you want to change the base?

2D Prostate Segmentation with Improved UNet - Student 47222610 #272

Uh oh!

Conversation

ChiaJouLu commented Nov 2, 2025

Project Summary

Implementation

Key Features:

Files:

Results

Uh oh!

ChiaJouLu commented Nov 3, 2025

Important Update: Corrected Class Labels

Uh oh!

wangzhaomxy commented Nov 21, 2025

Uh oh!

gayanku commented Nov 24, 2025

Marking

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants