{"id":25239708,"url":"https://github.com/akimuddinshaikh/master-s-research-project","last_synced_at":"2026-05-05T12:32:12.416Z","repository":{"id":275690981,"uuid":"926884585","full_name":"Akimuddinshaikh/Master-s-Research-Project","owner":"Akimuddinshaikh","description":"A hybrid approach combining texture-based (GLCM) and deep learning (ResNet50) features with unsupervised clustering and supervised classification for detecting liver diseases. Achieved 99%-100% accuracy using SVM, XGBoost, and Random Forest on pseudo-labeled medical imaging datasets","archived":false,"fork":false,"pushed_at":"2025-02-04T02:54:31.000Z","size":9924,"stargazers_count":0,"open_issues_count":0,"forks_count":0,"subscribers_count":1,"default_branch":"main","last_synced_at":"2025-04-05T20:09:47.636Z","etag":null,"topics":["agglomerative-clustering","dbscan-clustering","glcm","kmeans-clustering","opencv","python","random-forest","resnet-50","svm-model","tensorflow","xgboost-model"],"latest_commit_sha":null,"homepage":"","language":null,"has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/Akimuddinshaikh.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2025-02-04T02:49:41.000Z","updated_at":"2025-02-04T02:56:03.000Z","dependencies_parsed_at":"2025-02-04T03:37:04.558Z","dependency_job_id":null,"html_url":"https://github.com/Akimuddinshaikh/Master-s-Research-Project","commit_stats":null,"previous_names":["akimuddinshaikh/master-s-research-project"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Akimuddinshaikh%2FMaster-s-Research-Project","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Akimuddinshaikh%2FMaster-s-Research-Project/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Akimuddinshaikh%2FMaster-s-Research-Project/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/Akimuddinshaikh%2FMaster-s-Research-Project/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/Akimuddinshaikh","download_url":"https://codeload.github.com/Akimuddinshaikh/Master-s-Research-Project/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":247393572,"owners_count":20931813,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["agglomerative-clustering","dbscan-clustering","glcm","kmeans-clustering","opencv","python","random-forest","resnet-50","svm-model","tensorflow","xgboost-model"],"created_at":"2025-02-11T18:59:56.577Z","updated_at":"2026-05-05T12:32:12.380Z","avatar_url":"https://github.com/Akimuddinshaikh.png","language":null,"funding_links":[],"categories":[],"sub_categories":[],"readme":"Analysis of Scalable and Efficient Approaches for Liver Disease Detection\nA Hybrid Approach Using Texture-Based Features \u0026 Deep Learning\n\n📌 Author: Akimuddin Aslam Shaikh\n📍 Institution: National College of Ireland\n\n\n📌 Project Overview\nLiver disease detection using machine learning and deep learning techniques has gained attention for its potential to improve diagnostic accuracy and patient outcomes. This study introduces a unique hybrid approach by combining texture-based Gray-Level Co-occurrence Matrix (GLCM) features with high-dimensional deep learning features extracted from ResNet50.\n\nThe research focuses on leveraging unlabeled medical imaging datasets by integrating unsupervised clustering (KMeans, Agglomerative Clustering) with supervised classification (Random Forest, SVM, XGBoost). The results show a scalable and efficient methodology for detecting liver diseases with high precision.\n\n🔍 Key Highlights\n✅ Extracted GLCM features and ResNet50 deep learning features from medical imaging datasets.\n✅ Implemented unsupervised clustering (KMeans \u0026 Agglomerative Clustering) to generate pseudo-labels.\n✅ Trained classifiers (Random Forest, SVM, XGBoost) using extracted feature sets.\n✅ Combined features (GLCM + ResNet50) with cluster labels for improved classification accuracy.\n✅ Achieved near-perfect accuracy (99%-100%) in specific combinations but addressed overfitting concerns.\n\n📊 Machine Learning Techniques \u0026 Results\nFeature Set\tClustering Method\tClassifier\tAccuracy\nResNet50 Features\tKMeans\tRF/SVM/XGBoost\t~95%\nResNet50 Features\tAgglomerative Clustering\tRF/SVM/XGBoost\t99%-100%\nGLCM Features\tKMeans\tRF/SVM/XGBoost\t~98%\nGLCM Features\tAgglomerative Clustering\tRF/SVM/XGBoost\t100% (Potential Overfitting)\nGLCM + ResNet50 Features\tKMeans\tSVM\t~96%\nGLCM + ResNet50 Features\tAgglomerative Clustering\tSVM\t99%\n🔹 Key Finding: Combining ResNet50 deep learning features with Agglomerative Clustering labels yielded the highest classification accuracy (99%-100%), significantly outperforming other combinations.\n🔹 Overfitting Observed: Results exceeding 99% accuracy in real-world medical datasets indicate a potential overfitting issue, requiring further validation.\n🔹 Pseudo-Labeling Approach Improved Performance: Integrating pseudo-labeling from clustering techniques enhanced classification accuracy.\n\n🛠️ Technologies \u0026 Tools Used\nPython 🐍\nScikit-Learn (Machine Learning Models)\nTensorFlow/Keras (Deep Learning \u0026 ResNet50)\nOpenCV (GLCM Feature Extraction)\nSciPy \u0026 NumPy (Clustering \u0026 Statistical Analysis)\nMatplotlib \u0026 Seaborn (Data Visualization)\n📌 Results \u0026 Insights\n✅ GLCM + ResNet50 features provide a robust framework for liver disease detection.\n✅ Agglomerative Clustering outperformed KMeans in feature representation.\n✅ Pseudo-labeling from clustering significantly improved classification accuracy.\n✅ The study presents a scalable, efficient, and high-precision methodology for medical imaging analysis.\n\n📌 Future Work\n✔ Validate the approach on larger and more diverse medical imaging datasets.\n✔ Implement Transfer Learning with advanced architectures (e.g., Vision Transformers).\n✔ Optimize feature selection to mitigate overfitting risks.\n✔ Develop an explainable AI (XAI) model to enhance medical interpretability.\n\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fakimuddinshaikh%2Fmaster-s-research-project","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fakimuddinshaikh%2Fmaster-s-research-project","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fakimuddinshaikh%2Fmaster-s-research-project/lists"}