Deep learning approach to identify abnormalities in blood cell images
F. Mary Magdalene Jane, V. Pream Sudha, A. C. Soundarraj · International Journal of Health Sciences · 2022
Blood cell imaging provides crucial diagnostic information about a person's atypical problems. Due to their multi-level structures, Deep Learning models aid in extracting complicated insights from input images. The power of automatic feature extraction makes Convolutional Neural Networks (CNN) the most widely used deep learning approach for blood cell categorization. Blood cells are classified using a recurrent neural network (RNN), which captures long-term relationships between elements. The possibility of CNN-RNN-based models for blood cell image classification is investigated in this work. The results prove that CNN-LSTM based model outperforms in classifying the blood cell images with an accuracy of 92.1%.