Predicting in-hospital mortality is a critical problem in the medical field, aiding doctors in using appropriate treatment methods. This study focuses on evaluating and comparing various models for predicting patient mortality using the MIMIC-III database. The models examined include Logistic Regression, Random Forest, Convolutional Neural Networks (CNN), Vision Transformer (ViT), and Time-Series Representation Learning via Temporal and Contextual Contrasting (TS-TCC). The results indicate that most models achieved high accuracy, with ViT standing out due to its short training time while maintaining accuracy in comparison with other models. However, the TS-TCC model proved unsuitable in this study. The research highlights the potential for applying image processing models to time series data in medicine, opening up promising avenues for future research.