Brain tumor detection and classification from MRI images is a challenging task. Early and accurate diagnosis are essential for selecting appropriate treatment plans and improving patient outcomes. Despite significant advances in deep learning for medical image recognition, comprehensive comparative analyses of brain tumor classification models, particularly regarding ensemble optimization, remain limited. This paper uses four state-of-the-art deep learning frameworks, namely EfficientNetB4, MobileNetV3, MobileNetV2, and EfficientNetB0, to classify brain MRI images into four categories: Glioma, Meningioma, Pituitary tumor, and Normal. It employs a two-phase transfer learning approach, followed by 5-fold cross-validation on 875 MRI images. A unified experimental framework is employed, incorporating a two-phase transfer learning approach, consistent preprocessing, and a rigorous evaluation protocol with 5-fold cross-validation and an independent test set to prevent data leakage. Both full and selective ensemble strategies are examined to improve the robustness and stability. The models are evaluated using accuracy, precision, recall, F-1 score, confusion matrices, and accuracy curves, and statistical validation using McNemar’s test. MobileNetV3 achieves the highest test accuracy of 98.76%, followed by EfficientNetB4 (97.89%) and EfficientNetB0 (93.48%). MobileNetV2 performs significantly worse, with an accuracy of less than 80%. The selective ensemble technique (which uses the best models) attains the highest accuracy of 92.97%, compared to the full ensemble (84.40%), which improves prediction robustness but does not surpass the best individual model in peak accuracy. Overall, it can be concluded that MobileNetV3 is the most suitable architecture for brain tumor classification, delivering high accuracy with minimal computational complexity. The selective ensemble approach also enhances performance, maintaining computational efficiency, emphasizing the importance of informed model selection in neuro-oncological image analysis and clinical decision-support systems.

Image quality assessment (IQA) is a process of measurement of the image quality using the evaluations of subjective value with the model of computation. The quality of the image can be calculated by using different types of method where each method works with using isolated features of image. One very renowned method is structural similarity index (SSIM) which measured the quality of image comparing structure of image and the structure stage is obtained from pixel-based stage. FSIM (Feature Similarity Index) measured image quality using low level feature and Gradient magnitude (GM) act as primary feature of image. In this work, a novel MFSIM (Moderate Feature Similarity Index) is introduced which work with full reference IQA, HVS (Human Visual System) and low-level feature of images. In MFSIM the Phase Congruency (PC) is used as primary feature where the PC is dimensionless contrast invariant. In the moderated FSIM the Gradient Magnitude (GM) of the image is considered as the feature of secondary. For application IQA, we applied into segmented image with original image using MRI images. The distortion level of the segmented image is calculated using different image quality index measurement techniques. The image can be used in numerous purposes and the quality of image is distorted for different reason. There are lots of applications where noise less of perfect image is used for getting exact result. So it is very important to find out the distortion level of image. For instance during the segmentation of MRI image for brain tumor detection, the exactness of image need to calculate so that the brain tumor can be find out accurately. So the main purpose of this research work is to introduce a new image quality index and find out the brain tumor and the segmented image quality.