A hybrid approach with modified RCNN-based ROI identification and spot detect-X fusion strategy for DNA microarray analysis

Abstract

As the fourth most frequent type of cancer in both men and women, lymphoid malignancies pose a significant threat to healthcare. This type of cancer has also contributed to increased mortality rates due to the inaccurate detection of diseases. To address the limitations of existing approaches, a novel method called a Hybrid Modified RCNN-based ROI Identification and Spot Detect-X Fusion Strategy for DNA microarray analysis is proposed. Initially, the data undergoes pre-processing using various techniques such as Image Augmentation, Quantile Normalization, Background Correction, and Noise Reduction with an Improved Median Filter. ROI identification is performed using the modified RCNN. Subsequently, features are extracted, and feature selection is based on the Hybrid Anaconda-based Serval Optimization. Finally, diseases are classified using the Spot Detect-X Fusion Strategy. The proposed model achieved an accuracy of 98.98% when implemented in a MATLAB platform. Therefore, the recommended model demonstrates superior performance compared to existing methods.