Liver tumors rank as the fifth most prevalent cancer in men and the ninth in women, as reported in the 2018 Global Cancer Statistics. Established diagnostic methods such as Computed Tomography (CT), Magnetic Resonance Imaging (MRI), and ultrasound involve invasive and time-intensive procedures. This study proposes leveraging deep learning, specifically the ResUNet model, for early tumor detection, offering a more efficient alternative to traditional approaches. In contrast to previous research using dual cascaded CNNs, the ResUNet employs residential blocks, demonstrating its effectiveness on the 3D-IRCADb01 dataset derived from CT slices of liver tumor patients. Results indicate an impressive True Value Accuracy of approximately 99% and an F1 score performance of about 95%. This innovative method holds potential for early and accurate liver tumor diagnosis, contributing significantly to the biotechnology sector and potentially saving lives.