Liver cirrhosis, a widely prevalent pathology, is characterized by a insidious progression, often asymptomatic until an advanced stage. Despite medical advances, this condition remains potentially fatal. Early detection of hepatic fibrosis is crucial to improve survival rates. However, current screening methods have significant constraints: ultrasound images, often with insufficient resolution, can complicate interpretation and lead to misdiagnoses; moreover, non-invasive alternatives such as FibroTest or FibroScan, while accurate, are expensive and less accessible. Furthermore, the delays associated with shipping samples abroad for detailed analysis can be detrimental, with speed being a key factor in managing this disease. In this context, the application of artificial intelligence in the diagnosis and detection of liver fibrosis has become indispensable. With this in mind, three new models have been developed: an EfficientNet model designed for precise identification of ultrasound images, a model based on Vision Transformer technology for classifying livers into ‘healthy’ or ‘fibrotic’ categories, and an occlusion model to identify important areas on which our vision model relies to classify our images as healthy or fibrotic livers.