VGG16-PCA-PB3C: A hybrid PB3C and deep neural network based approach for leukemia detection

In recent decades, the leukemia detection using deep learning and convolutional neural networks have shown to be more effective as the main and the important features are retrieved by these methods. Early detection of leukemia is crucial for timely treatment and improved outcomes. Although leukemia detection using convolutional neural networks has been demonstrated to be effective, there is still potential for improvement in terms of accuracy and efficiency of the model. The main objective of this paper is to present a novel Hybrid Deep Neural Network, termed Visual Geometry Group16-Principal Component Analysis-Parallel Big Bang Big Crunch (VGG16-PCA-PB3C), for the accurate and timely detection of leukemia. This approach integrates Visual Geometry Group (VGG16) for robust feature extraction, Principal Component Analysis (PCA) for effective dimensionality reduction by minimizing the amount of training time and enhancing the accuracy, and the innovative Parallel Big-Bang-Big-Crunch (PB3C) optimization algorithm to address computational complexities during training. The integration of these techniques aims to improve the accuracy and effectiveness of leukemia detection. We evaluated our proposed framework on Classification of Normal versus Malignant Cells (CNMC) dataset which has 15,384 blood smear images. For the leukemia detection, VGG16-PCA-PB3C achieved an accuracy of 95%, and a precision of 94.0%. It is demonstrated by the experimental results that the classifier's performance has increased, and the suggested model performed better as compared to the other popular image classification approaches. The proposed framework can be applied to create accurate and efficient models for leukemia detection, which can assist medical professionals in the diagnosis of leukemia.