E-DCNN: Ensemble-Based Deep CNN Model for Acute Lymphoblastic Leukaemia (ALL) Classifier Using Microscopic Blood Smear Images

A malignant disease called acute lymphoblastic leukemia (ALL) is exemplified by an anomalous build-up of immature cells in the blood. Early disease detection is closely correlated with effective ALL treatment. The manual examination of stained blood smear microscopy pictures used in current practice in diagnosing ALL is overwhelming in time usage and prone to error. Recently, a potent tool for aiding doctors in making medical decisions has emerged: Artificial intelligence-oriented manual-dependent diagnosis. To automatically recognize ALL in blood pictures, various computer-aided diagnostic techniques have been created. To detect ALL in stained blood smear microscopy pictures, a new Ensemble-based Convolutional Neural Network, E-DCNN model based on Bayesian optimization, is presented in this paper. The proposed E-DCNN model with its hyperparametric values is tailored to the imagery input data using the Ensemble-based Deep learning optimization approach to improve classification performance. An amalgam data repository is created by combining two open ALL datasets and is used to train and validate the suggested CNN. Data augmentation is utilized to enhance the hybrid image set and improve categorization performance. On the test set, the ensemble-based classifier-derived ideal CNN model performed better at image-based ALL classification. The results of this investigation show that the suggested E-DCNN is superior to previous optimized deep learning ALL classification methods.