AI-DRIVEN Novel Approach for Liver Cancer Screening and Prediction Using Cascaded Fully Convolutional Neural Network

Abstract

In experimental analysis and computer-aided design sustain scheme, segmentation of cell liver and hepatic lesions by an automated method is a significant step for studying the biomarkers characteristics in experimental analysis and computer-aided design sustain scheme. Patient to patient, the change in lesion type is dependent on the size, imaging equipment (such as the setting dissimilarity approach), and timing of the lesion, all of which are different. With practical approaches, it is difficult to determine the stages of liver cancer based on the segmentation of lesion patterns. Based on the training accuracy rate, the present algorithm confronts a number of obstacles in some domains. The suggested work proposes a system for automatically detecting liver tumours and lesions in magnetic resonance imaging of the abdomen pictures by using 3D affine invariant and shape parameterization approaches, as well as the results of this study. This point-to-point parameterization addresses the frequent issues associated with concave surfaces by establishing a standard model level for the organ’s surface throughout the modelling process. Initially, the geodesic active contour analysis approach is used to separate the liver area from the rest of the body. The proposal is as follows: It is possible to minimise the error rate during the training operations, which are carried out using Cascaded Fully Convolutional Neural Networks (CFCNs) using the input of the segmented tumour area. Liver segmentation may help to reduce the error rate during the training procedures. The stage analysis of the data sets, which are comprised of training and testing pictures, is used to get the findings and validate their validity. The accuracy attained by the Cascaded Fully Convolutional Neural Network (CFCN) for the liver tumour analysis is 94.21 percent, with a calculation time of less than 90 seconds per volume for the liver tumour analysis. The results of the trials show that the total accuracy rate of the training and testing procedure is 93.85 percent in the various volumes of 3DIRCAD datasets tested.

An Innovative Machine Learning Approach for Classifying ECG Signals in Healthcare Devices

“…For performing classification, training and test datasets are obtained from MIT-BIH database, which has both normal and patient datasets [ 24 ]. Nearly 80% of data are chosen for training and 20% was considered for testing.…”

Two-Way Feature Extraction for Speech Emotion Recognition Using Deep Learning

“…Another effective and popular way to extract emotions from speech was to use CNNs [ 7 , 8 ] and Deep CNNs [ 9 , 10 ]. The authors in [ 11 ] used the Deep CNNs to approach the task.…”

Augmented Reality-Centered Position Navigation for Wearable Devices with Machine Learning Techniques

“…The passive tags are not provided with any energy, and they obtain power from the readers. However, the other tags are provided with an energy storage component [ 26 ].…”

“…Bluetooth beacons have greater accuracy compared with Wi-Fi access points, but the downside to this accuracy is that it is directly proportional to the number of beacons used. Mounting beacons over a small area such as an office is easy and cost-effective, but when it comes to a larger structure such as universities, it becomes impractical [ 26 ]. The problem with the Wi-Fi access points is not only its accuracy but also the fact where failing to establish a connection to the device or a delay in the connection may result in errors and false mapping of the user's position.…”

“…Augmented reality (AR) specializes in overlaying the real-world environment with virtually generated objects. The goal of AR is to provide the most clear and accurate representations where the user finds it hard to differentiate the virtual augmentations that are being applied [ 26 ]. AR is employed in various applications, including game development, military training, simulations, design, and entertainment [ 28 ].…”