AI SYSTEM ON CHEST CTX-RAY IMAGE CLASSIFICATION USING MTL AND SGDM OPTIMIZATION TRAINING

• Main Authors: Chacko, Susamma, R., Krishna Priya, Bhupathyraaj, Mullaicharam, Ahamed, Suhail, Bimani, Ali Al
• Format: Patent
• Language: English
• Published: 17.06.2021
• Subjects: CALCULATING; COMPUTER SYSTEMS BASED ON SPECIFIC COMPUTATIONAL MODELS; COMPUTING; COUNTING; HANDLING RECORD CARRIERS; HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATIONTECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING ORPROCESSING OF MEDICAL OR HEALTHCARE DATA; INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTEDFOR SPECIFIC APPLICATION FIELDS; PHYSICS; PRESENTATION OF DATA; RECOGNITION OF DATA; RECORD CARRIERS

Abstract Respiratory Failure has been a major source of anxiety for countries all over the world. The pandemic's accelerated spread has caused disruption and necessitated quick responses to mitigate the consequences. Many studies in a number of fields of science have begun to look at the issues in order to address them. Artificial Intelligence (AI) is one of the fields of research that has had a lot of success in solving problems in a variety of ways. This invention aimed to propose a CNN model to diagnose affected patients from chest CT/X-ray images. Modified Transfer Algorithm concept is used in the well-known models of CNN for image recognition namely VGG19 Net, Dense Net, and Res Net 50 and also on Ensemble algorithm to observe acceptable results. In this process of multiple modified transfers, the algorithm is used for the modified convolutional layers. The process of modification and training is performed by using SGDM optimization. The data set used will be either CT/X ray images. The larger the data set, the better will be the training. The modified Transfer learning will train the data set using the 3 CNNs mentioned and Modified Transfer Learning with Ensemble algorithm will train on CNN models like MatCony Net and Alex Net. It will get compared and evaluated with the testing and validation process. The training accuracy has crossed above 90%, thus the model is in an acceptable method. AT SYSTEM ON CHEST CTX-RAY IMAGE CLASSIFICATION USING MTL AND SGDM OPTIMIZATION TRAINING Diagram CTSCAN X-RAY SCAN 1. 3D IMAGE 1. 2D IMAGE 2. USED AS A PRIMARY SOURCE 2. USED PRIMARILY TO DETECT IN DIAGNOSING SOFT TISSUE AND CANCER AND PNEUMONIA AND ORGAN TO MONITOR BONES 3. POWERFUL THAN X-RAYS 3. WIDELY USED IN MEDICAL FIELD 4.36D DEGREE IMAGE 4. USE RADIATION TO PRODUCE Figure 1: Difference between X-Ray and CT scan.