Blood Cells Analysis

A project for analysing different types of blood cells in images.

The analysis of an image provided in the dataset can have three types of results:

1. Lymphocyte;
2. Neutrophil;
3. None of the above.

The model used for training and testing is the Naive Bayes Classification Model.

The chosen descriptor is the HuMoments descriptor.

Features:

• Region of interest (ROI) selection;
• Image noise reduction;
• Image sharpness improvement;
• Image quality enhancement (contrast, brightness and histogram equalization);
• ROI cropping and filter applying;
• Creating descriptors for images;
• Model training and testing;
• Model exportation;
• Object detection (with or without the usage of a slider).

© Muftić Belma, Lemeš Lamija & Krupalija Ehlimana

Faculty of Electrical Engineering Sarajevo, 2018.