Novel Ultrasonic Image Reconstruction and Deep Learning Model-Based Categorization for Accurate Liver Disease Diagnosis

• Published in: 2023 International Conference on Data Science and Network Security (ICDSNS) pp. 1 - 6
• Main Authors: S, Nijaguna G, Sureddy, Sneha, BinduMadhavi, Mohapatra, Dayanidhi, D, Parameshachari B
• Format: Conference Proceeding
• Language: English
• Published: IEEE 28.07.2023
• Subjects: Acoustics; Bariatric Surgery Patients; Deep Convolutional Neural Network; Deep learning; Fatty Liver Disease; Inceptionresnetv2; Liver diseases; Reliability; Sensitivity; Training; Ultrasonic Image Reconstruction; Ultrasonic imaging; Ultrasound Images
• DOI: 10.1109/ICDSNS58469.2023.10245601

The diagnosis of liver illnesses is critical for efficient treatment and management of the condition. By combining deep learning model-based classification with ultrasonic image reconstruction, the developed method aims to surpass standard deep learning techniques. This method's nine primary processes include pre-processing approaches, image reconstruction, liver area segmentation, feature extraction, deep learning model training, testing, and comparing model performance. The suggested approach was tested on patients with fatty liver disease by processing huge sets of ultrasound pictures and RF data frames. The InceptionResNetV2 CNN was used to detect fatty liver disease in these persons. The proposed approach has a sensitivity of 95.2 percent, an accuracy of 92.8 percent, and a specificity of 94.1 percent. By combining deep learning model-based classification with ultrasonic image reconstruction, the technology overcomes some of the limitations of classic deep learning methods. Medical ultrasound imaging is commonly used because it provides for the non-invasive, risk-free identification of liver disorders. Ultrasonic imaging is a sort of medical image diagnostic that assists doctors in diagnosing and treating a variety of ailments. This suggests that the suggested technique might considerably improve the precision and consistency with which liver disease is detected, resulting in better patient outcomes.