Topic recognition

[AshHarikrishna]

This pull request introduces the Topic Recognition module, which enables automated identification and classification of topics within the dataset.

extension for Monday 3 Nov, 4pm

Key Features

• End-to-end topic recognition pipeline, including preprocessing, model training, and evaluation.
• Integration with the existing data handling framework for seamless input/output compatibility.
• Initial experiments demonstrating topic detection accuracy and reliability.

###Key Components

• convert.py: Converts raw ISIC JSON annotations into binary label arrays (binary_labels.npy), simplifying lesion presence detection and preparing structured label data for YOLO training.
• prepare_yolo.py: Splits dataset into train/val/test sets and generates YOLO-format .txt labels for each image. Ensures balanced stratification across lesion types.
• dataset.py: Dynamically generates the data.yaml configuration file, specifying dataset paths, class names, and total class count for YOLOv8 training.
• modules.py (ISICDetector): Defines the core detector class built around YOLOv8. Handles training configuration, hyperparameter tuning, and inference logic for lesion detection.
• train.py: Training script that orchestrates YOLO training (calls get_data_yaml / dataset.py as needed), sets hyperparameters, and saves training logs and weights to runs/.
• predict.py: Loads the trained YOLO model and performs predictions on new images, saving visual detections and optional cropped regions for downstream classification.
• classify_part3.py (optional): Utilizes YOLO-generated lesion crops to train a lightweight CNN (e.g., ResNet18) for final lesion-type classification.

[shaivikaaaa]

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

YOLOv8 → Normal Difficulty

Category		Marks	Comments
Algorithm solves the problem	5	3	no plots are present, or any image generated by the model, only numerical values present in README
Implementation functions as intended	3	2	The pipeline might run, but there’s no evidence that predictions are correct or that outputs are usable
Good design	1	1
Commenting	1	1
Algorithm above Normal Difficulty	5	5
Algorithm is Hard difficulty	5	0	Normal Difficulty
Section IV : Max mark 15 from 20		12

• Discussion: None provided. This is needed.

Suggestions/Notes:

• no plots, no images are present. You were suppose to give some working evidence of your model
• Good job otherwise!

[wangzhaomxy]

47511891

[gayanku]

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.	-1
	Driver Script present.
	File structure present.
	Good Usage & Demo & Visualisation & Data usage.
	Module present.
	Commenting present.
	No Data leakage found.
	Difficulty : Normal. Normal Difficulty : Yolov8	-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