Search

Nowadays working in the medical imaging domain remains a big challenge, since the collecting such datasets is so complex, deep learning techniques have accomplished great success in a diversity of applications across many disciplines. In current work, we have combined two novel neural networks for classifying liver lesions of Hepatocellular carcinoma, Metastases, Hemangiomas, and Healthy tissues. On the first model called FireNet, we have introduced fire modules to reduce the model size and the number of parameters for quick classification. The first part will be in charge of extracting spatial feature information. The second model called Modified Long Short-Term Memory (MLstm), the features extracted by the FireNet are then used for temporal information and prediction with a new loss function called G-loss. In order to improve the proposed FireNet-MLstm model, a new bias was added through the forget gate. The activation function and the hyperbolic tangent were also added, which increased prediction accuracy. The dataset used in this study was collected by searching for the medical records with HCC, MET, HEM, and Healthy tissues in Jiangbin Hospital, the affiliated hospital of Jiangsu University from 2015 to 2018, with 120 patients, 30 patients with one or multiple Hcc, 26 patients with one or multiple Hem, 23 patients with one or multiple Met and 41 Healthy with non-lesion. The final classification accuracy of the proposed FireNet-MLstm model was 91.2%. [ABSTRACT FROM AUTHOR]

ObjectiveThe purpose of this article is to review the imaging features and Liver Imaging Reporting and Data System (LI-RADS) categorization of benign and likely benign entities, including typical cirrhotic nodules, distinctive nodular observations, and benign entities that may simulate hepatocellular carcinoma.ConclusionLI-RADS is a system of standardized criteria for interpreting liver CT and MR images of patients at risk of hepatocellular carcinoma. Most of the observations in these patients are not malignant. With the development of fibrosis and cirrhosis, these benign entities may take on an altered appearance.

In recent years, deep learning techniques, and in particular convolutional neural networks (CNNs) methods have demonstrated a superior performance in image classification and visual object recognition. In this work, we propose a classification of four types of liver lesions, namely, hepatocellular carcinoma, metastases, hemangiomas, and healthy tissues using convolutional neural networks with a succinct model called FireNet. We improved speed for quick classification and decreased the model size and the number of parameters by using fire modules from SqueezeNet. We have used bypass connection by adding it around Fire modules for learning a residual function between input and output, and to solve the vanishing gradient problem. We have proposed a new Particle Swarm Optimization (NPSO) to optimize the network parameters in order to further boost the performance of the proposed FireNet. The experimental results show that the parameters of FireNet are 9.5 times smaller than GoogLeNet, 51.6 times smaller than AlexNet, and 75.8 smaller than ResNet. The size of FireNet is reduced 16.6 times smaller than GoogLeNet, 75 times smaller than AlexNet and 76.6 times smaller than ResNet. The final accuracy of our proposed FireNet model was 89.2%. [ABSTRACT FROM AUTHOR]

Objective: The purpose of this article is to review the imaging features and Liver Imaging Reporting and Data System (LI-RADS) categorization of benign and likely benign entities, including typical cirrhotic nodules, distinctive nodular observations, and benign entities that may simulate hepatocellular carcinoma., Conclusion: LI-RADS is a system of standardized criteria for interpreting liver CT and MR images of patients at risk of hepatocellular carcinoma. Most of the observations in these patients are not malignant. With the development of fibrosis and cirrhosis, these benign entities may take on an altered appearance.