Bone Cancer Detection Using Feature Extraction Based Machine Learning Model.

Area of Science:

• Medical Imaging Analysis
• Computational Pathology
• Machine Learning in Oncology

Background:

• Bone cancer poses a significant health risk, often leading to fatal outcomes.
• Current diagnostic methods like X-ray, MRI, and CT scans are manual, time-consuming, and require specialized expertise.
• Distinguishing cancerous from healthy bone tissue can be challenging due to similar morphological characteristics in medical images.

Purpose of the Study:

• To develop an automated system for accurate classification of cancerous and healthy bone tissues.
• To address the limitations of manual image analysis in bone cancer diagnosis.
• To compare the effectiveness of different feature sets and machine learning models for bone cancer identification.

Main Methods:

• Evaluation of various edge detection algorithms to identify the most suitable one for bone image analysis.
• Preparation of two distinct feature sets: one incorporating Histogram of Oriented Gradients (HOG) features and another without.
• Utilizing Support Vector Machine (SVM) and Random Forest machine learning models to classify bone tissue based on the prepared feature sets.

Main Results:

• The feature set incorporating HOG demonstrated superior performance compared to the set without HOG across both machine learning models.
• The SVM model trained with the HOG feature set achieved a high F1-score of 0.92.
• The Random Forest model achieved an F1-score of 0.77 when trained with the HOG feature set.

Conclusions:

• The integration of HOG features significantly enhances the accuracy of automated bone cancer detection systems.
• The SVM model, combined with HOG features, proves to be a highly effective tool for classifying cancerous bone tissue.
• This automated approach offers a promising alternative to manual diagnostic methods, potentially improving efficiency and accuracy in bone cancer identification.