IMAGE CLASSIFICATION OF WHITE BLOOD CELLS USING DEEP LEARNING
https://github.com/kittiekat/WBC_Classifier/archive/refs/heads/main.zip
The data set utilized is images of white blood cells from regular peripheral blood samples known as Raabin-WBC posted in 2021. The dataset carries approximately 30000 WBCs and artefacts (colour spots). “To reassure accurate data, a considerable wide variety of cells had been categorized via way of means of specialists, and the ground truth of nucleus and cytoplasm had been extracted (approximately 1145), as well”. In 5 distinct files, 5 different kinds of cells had been stored. This blood dataset is appropriate for building a picture classifier that classifies blood photos into relevant categories employing a kind of approaches associated methodologies. All these films were stained through Giemsa technique. the conventional peripheral blood smears are taken exploitation the camera phone of Samsung Galaxy S5 and therefore the magnifier of Mt. Olympus CX18. Also, the CML slide has been imaged utilizing an LG G3 camera phone at the side of a microscope of Zeiss brand. It's value noting that the pictures have all been enamored a magnification of a hundred.
CNN ~ 0.94
ResNet50 ~ 0.98
AlexNet ~ 0.91
InceptionV3 ~ 0.94
VGG16 ~ 0.98