In recent years, kidney disease has emerged as a significant public health concern, as indicated by objective assessments of renal health and damages. Classification of kidney diseases is tedious when done manually. An effective machine learning model will always help in auto diagnosis of kidney disease from CT scan. In this study, an efficient and composite model for kidney CT scan image identification and classification is presented. A normalized approach is used to solve the issues present in the existing studies. The first step is pre-processing the photos, which consists of rescaling, converting them to grayscale, and removing noise by combining the Wiener and median filters. A Convolutional Neural Network is then used for extraction of features from the image. Aquila-optimized upgraded ResNet-101 model is used for classification. The model’s efficacy is evaluated in comparison to similar research. With an accuracy of 99.20%, specificity of 98.30%, F1-score of 98.45%, sensitivity of 98.10%, and precision of 98.03%, the suggested model has demonstrated its efficacy.