Article’s

Liver cancer is one of the leading causes of cancer-related deaths worldwide, largely due to delayed diagnosis and the complexity of identifying tumors at an early stage. Medical imaging techniques such as Computed Tomography (CT) and Magnetic Resonance Imaging (MRI) are widely used for liver examination; however, manual interpretation of these images is time-consuming and highly dependent on clinical expertise. To address these challenges, this study presents a Deep Learning Based Liver Tumor Detection System that utilizes Convolutional Neural Networks (CNNs) for automated tumor identification from medical scan images. The proposed framework employs a pre-trained EfficientNet model for feature extraction and image classification, enabling accurate differentiation between normal and tumor-affected liver images. To improve interpretability, Grad-CAM visualization is incorporated to highlight image regions that influence the model’s predictions. A Flask-based web application is developed to provide secure user access, image upload functionality, prediction visualization, and performance analysis. Experimental results demonstrate that the system can effectively detect liver tumors and provide reliable classification outcomes. The proposed solution serves as an intelligent decision-support tool that assists healthcare professionals in early diagnosis, reduces diagnostic workload, and enhances clinical decision-making efficiency.