This work presents detection of liver tumours in computed tomography (CT) images based on image processing and deep learning. The existing systems for liver tumour detection are based upon support vector machine (SVM) and random forest approach, which are less interpretable, computationally expensive, and require more parameters. The proposed approach commences with a comprehensive pre-processing stage encompassing resizing, histogram equalization, and bilateral filtering for noise reduction. In addition, K-Means image-based segmentation has been employed to extract the CT images’ features effectively. The pivotal element of this approach is based on the utilization of convolutional neural networks (CNNs), a potent deep learning tool, to achieve accurate tumour identification by classifying them as benign or malignant. By capitalizing on the strengths of deep learning, specifically CNNs, this method aims to significantly elevate the precision of liver tumour detection within CT scans.