Lymphoma Classification Using Convolutional Neural Network With Dense Attention Learning Strategy

Lymphoma is a group of blood cancers that affect the lymphatic system, which is an important part of the body's immune system. The classification of lymphoma cancer depends on the blood cells, namely T‐cells and B‐cells. The categorization of lymphoma is challenging due to the presence of solid tumors and leukemia. However, using histopathological images for lymphoma classification is challenging due to inter‐ and intra‐class similarities in subtle variations and tissue structure, which reduce the model accuracy. Hence, this research proposes the Dense Attention Learning Strategy with Convolutional Neural Network (DALS‐CNN) for accurate lymphoma classification. The DALS method uses convolutions to efficiently manage correlations between the corrupted and uncorrupted regions of histopathological images. The min‐max normalization, Gaussian filtering technique, and data augmentation help to improve the image quality. A ResNet50 model is used to extract the features, which significantly improves the model's performance. The proposed DALS‐CNN method achieves an accuracy of 99.88% on the multicancer dataset when compared with the conventional techniques.