Breaking Barriers in Brain Tumor Detection with UWB Circular Monopole Antenna and Inception ResNet V2 in Microwave Imaging

The early detection of brain tumors is vital for improving survival rates and reducing risks associated with late-stage diagnosis. The current study addresses the limitations of existing imaging and diagnostic techniques by integrating microwave imaging using an ultra-wideband (UWB) circular monopole patch antenna with a deep learning (DL)-based classification framework. The antenna, designed with a defected ground structure (DGS), operates at dual resonance frequencies of 4.7 GHz and 7.58 GHz, achieving a return loss of -48.3 dB, a directivity of 6.79 dB, and a gain of 5.62 dB. Microwave images, generated through an innovative nine-antenna array system, provided a non-invasive, non-ionizing diagnostic solution. These images were analyzed using the Inception-ResNet v2 model, which combines multi-scale feature extraction with residual connections for effective brain tumor detection. A dataset of 1,282 images, was used for training and evaluation. The proposed framework demonstrated high reliability, achieving an accuracy of 98.4%, a precision of 97.9%, and a recall of 98.2%. The study’s outcomes highlight the potential of combining UWB microwave imaging with DL to revolutionize brain tumor diagnosis, offering a faster, safer, and highly accurate alternative to existing modalities, thereby meeting the pressing need for improved diagnostic capabilities in clinical settings.