Reduction of noise in multimodal brain images using adaptive filtering techniques
N. Thenmoezhi, B. Perumal, Adidela Rajya Lakshmi, M. Pallikonda Rajasekaran, Kottaimalai Ramaraj, Arunprasath Thiyagarajan · 2025
An autonomous brain tumour classification system was built by combining an Adaptive Filter with a Neural Network using imaging and processing techniques. The traditional method for categorizing brain MRI and detecting malignancies is human evaluation. Operator-assisted methods for classification are not feasible and consistent for large amounts of data. Medical resonance pictures have noise from operator error that can lead to significant categorization errors. Symbolic logic, adaptive filters, and rapid neural networks are three AI techniques that have demonstrated a lot of promise in this subject. Therefore, the adaptive filter is applied to the PETS can image in this paper, and a neural network is used for the intended purposes. A neural network classifier was used to meet the requirements. Two phases of classification were carried out: the probabilistic Neural Network (NN) and GLCM. Classification accuracy and coaching performance were used to assess the NN classifier’s performance. Neural networks are potentially the most effective technique for classifying cancers since they provide quick and precise classification. However, brain tumour detection at an early stage is a challenging endeavor. Due to the inaccurate segmentation results caused by the tumour’s soft edges in the PET image. In this article, the area was also estimated following the fuzzy c mean clustering method and probabilistic neural network, and the adaptive filtering technique was utilized for the detection and diagnosis of the brain tumour.