Intelligent Ultra-Light Deep Learning Model for Multi-Class Brain Tumor Detection

The diagnosis and surgical resection using Magnetic Resonance (MR) images in brain tumors is a challenging task to minimize the neurological defects after surgery owing non-linear nature of size, shape, textural variation. Radiologists, clinical experts, surgeons examine MRI scans available methods, which are tedious, error-prone, time-consuming, still exhibit positional accuracy up 2–3 mm, very high case cells. In this context, we propose an automated Ultra-Light Brain Tumor Detection (UL-BTD) system based on novel Deep Learning Architecture (UL-DLA) for deep features, integrated with highly distinctive extracted by Gray Level Co-occurrence Matrix (GLCM). It forms Hybrid Feature Space (HFS), used tumor detection Support Vector Machine (SVM), culminating prediction optimum false negatives limited network size fit within average GPU resources modern PC system. objective study categorize multi-class publicly datasets minimum time thus real-time can be carried out without compromising accuracy. Our proposed framework includes sensitivity analysis image One-versus-All One-versus-One coding schemes stringent efforts assess complexity reliability performance K-fold cross-validation as part evaluation protocol. best generalization achieved SVM has rate 99.23% (99.18%, 98.86%, 99.67%), F-measure 0.99 (0.99, 0.98, 0.99) (glioma, meningioma, pituitary tumors), respectively. results have been found improve state-of-the-art (97.30%) 2%, indicating that exhibits capability translation hospitals during applications. method needs 11.69 ms compared 15 earlier detect test any dedicated hardware providing route desktop application surgery.