Deep Learning Technique to Detect and Diagnose the Anomalous in Kidney

• Published in: 2023 International Conference on Innovative Computing, Intelligent Communication and Smart Electrical Systems (ICSES) pp. 1 - 6
• Main Authors: Sri, Vemu Santhi, Jothi Lakshmi, G. R.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 14.12.2023
• Subjects: Brain modeling; Computational modeling; Computer Tomography (CT); Deep learning; Kidney; Pre-trained model; Solid modeling; Stroke (medical condition); VGG-16 and InceptionV3 model; X-ray imaging
• DOI: 10.1109/ICSES60034.2023.10465442

Several diseases that have had a impact on our society is kidney abnormalities, which rank as such ninth highest prevalent malignancy both in men and women globally. Premature KT identification has substantial advantages in terms of decreasing mortality rates, generating precautionary measures that lessen consequences, and eradicating the tumor. Deep learning (DL) automatic recognition techniques can speed up diagnostic, increase test precision, lower performance expenditures, and lighten the stress of physicians as contrasted to the laborious and time-consuming conventional diagnostic testing. This study established a deep ensemble model for identifying and categorizing different types of kidney ailments, including cyst, stone, tumor, and normal using computer tomography photographs. In addition, while executing layers in a neural network, the testing accuracy of the deep based pre-trained model reached a higher accuracy of 97.8% with 0.1 ms loss. With hybrid and other algorithms without normalizing pictures, several studies looked at how well 95.7% of kidney problems could be identified. However, the primary goal of our effort was to normalize the pictures in order to greatly increase computing efficiency and convergence speed.