Particle Swarm Optimization Technique to Detect Brain Tumor using YOLO Technology

Abstract

The purpose of this study is to improve brain tumor detection through the strategic application of deep learning technologies. Early and accurate diagnosis of brain tumors is crucial for improving patient outcomes. For automated brain tumor diagnosis in MRI scans, this paper presents a novel synergy between particle swarm optimization (PSO) and the YOLOv7 deep learning system. The PSO technique is used for exact tumor region-of-interest (ROI) segmentation using the BRaTS21 dataset, a rigorously chosen benchmark from Kaggle. The tailored ROI is subsequently fed into the YOLOv7 architecture, allowing for accurate tumor location and classification. This integration takes advantage of PSO's powerful segmentation capabilities as well as YOLOv7's enhanced detection performance. The YOLOv7 model was trained particularly on the PSO-segmented BRATS21 data to validate the proposed architecture. Metrics such as as precision, recall, mean average precision (mAP), and F1 score are utilized to evaluate our system. In comparison to existing methodologies, the suggested solution outperforms them in terms of accuracy and processing speed. When compared to previous models, the PSO-YOLOv7 system displayed significant improvement, achieving higher accuracy and faster processing times. This integrated technique has enormous clinical application potential. The PSO-YOLOv7 framework, by accurately segmenting and classifying tumors, can pave the way for earlier diagnoses, better treatment decisions, and, ultimately, better patient outcomes.