“Integrated Hybrid Deep Learning Approach for Precise Blood Cancer Detection”

In this innovative endeavour, we propose an advanced method for the early identification and categorization of blood cancer, also known as leukaemia, utilizing cutting-edge deep learning techniques. Our project focuses on leveraging the MobileNetV2 architecture implemented in Python to achieve remarkable accuracy in diagnosing Acute Lymphoblastic Leukaemia (ALL) through peripheral blood smear (PBS) images. The dataset employed comprises four distinct classes, Benign, early Pre-B, Pre-B, and Pro-B cells, comprising a total of 3242 PBS images, exhibit a critical importance in accurate cancer screening. Precise diagnosis at this phase is crucial for effective early detection of malignancies. The deep learning model we’ve developed, based on MobileNetV2, exhibits exceptional performance with an impressive training accuracy of 98.00%. Its robustness is evident through a validation accuracy of 96.00%, showcasing its ability to effectively discern various blood cancer cell types. Ultimately, this advancement aims to improve patient outcomes by facilitating early detection and classification. The integration of MobileNetV2 and Python not only ensures efficiency but also makes our solution readily accessible to healthcare professionals, promising improved cancer screening and diagnosis.