Title: Precision Enhancement of Brain Tumors in MRI Scans: A Hybrid EADTV and Bilateral Filtering
Authors: Mahmoud Ibrahim AbuJalambo
Volume: 10
Issue: 1
Pages: 106-118
Publication Date: 2026/01/28
Abstract:
Magnetic Resonance Imaging (MRI) is a critical tool in the early detection and diagnosis of brain tumors. However, inherent noise, low contrast, and weak tumor boundaries often hinder the effectiveness of automated diagnostic systems. In this study, we propose a comprehensive and adaptive image enhancement and tumor localization pipeline tailored for brain MRI images. The methodology integrates multiple enhancement techniques-most notably, the Edge-Adaptive Directional Total Variation (EADTV) model and Bilateral Filtering-to improve the visual quality and clarity of tumor regions. The proposed approach begins with preprocessing stages, including contrast enhancement using Contrast Limited Adaptive Histogram Equalization (CLAHE) and noise reduction through Non-Local Means and Bilateral Filtering. Tumor regions are initially detected via K-Means clustering, which provides a rough but effective mask of the suspected lesion area. This mask is then used to isolate the tumor region from the original image. The isolated tumor is subsequently subjected to the EADTV model and anisotropic diffusion filtering to suppress residual noise while preserving edge details. To ensure the enhanced tumor maintains visual consistency with the original image, a selective histogram matching technique is applied exclusively to the tumor region before reintegration. This reintegration step improves the contrast between healthy and pathological tissues while avoiding global intensity distortions. Experimental results conducted on a small sample of brain MRI slices demonstrate the efficacy of the proposed technique. Peak Signal-to-Noise Ratio (PSNR) values reached up to 45.81, Structural Similarity Index Measure (SSIM) peaked at 0.9973, and Root Mean Square Error (RMSE) values remained low, confirming minimal distortion and high fidelity. These quantitative metrics, along with visual inspection, validate the potential of this approach for clinical applications, where precision in tumor boundary enhancement is crucial for diagnosis, surgical planning, and treatment monitoring.