Malaria Cell Detection using HOG
Detect malaria-infected cells from microscopy images using classical HOG features and an SVM classifier, with explainable results.
Technical Information
Project Type
Computer VisionMedical ImagingExplainable ML
Tech Stack
Python 3.8+OpenCVscikit-image (HOG)scikit-learn (SVM)NumPyMatplotlibJoblibStreamlitHugging Face Spaces
Topics / Modules
HOG Feature ExtractionSVM ClassificationExplainability
Links
Project Summary
A computer-vision app that classifies blood-cell microscopy images as infected or healthy using Histogram-of-Oriented-Gradients features and a Support Vector Machine – paired with visual explainability.
Skills Demonstrated
Image PreprocessingHOG FeaturesSVM ClassificationExplainable AIStreamlit Deployment
Solution
Images are preprocessed with OpenCV, described with HOG features via scikit-image, classified by an SVM, and the prediction is shown with an explainability view in a Streamlit app.
Approach
- Preprocess: Load and normalise cell microscopy images with OpenCV.
- Extract: Compute HOG feature descriptors.
- Classify: Train and apply an SVM classifier.
- Explain: Visualise the result with explainability.
- Deploy: Host the Streamlit app on Hugging Face Spaces.