Liver Cancer Detection in Histopathological Images: A Holographic Convolutional Neural Network Approach with Sea-Horse Optimization

Liver cancer is a major global contributor to morbidity, commonly arising alongside persistent liver disorders such as cirrhosis and hepatitis, which significantly increase the risk of its development. Survival in patients is largely determined by early and accurate diagnosis, but current diagnostic techniques and traditional deep learning approaches are prone to performance issues and instability. Going beyond these limitations, this paper introduces a holographic convolutional neural network with a superimposed sea-horse optimization algorithm (HoloCNN-SHO) as a robust liver cancer classifier. The approach entails advanced pre-processing, segmentation, and holographic encoding of the features to identify weak spatial–frequency features essential for histopathological detection. Evaluated on two benchmark datasets (LiTS and KMC Liver), the system demonstrated 99.9% accuracy with an extremely low false-positive rate of 0.2%. The findings emphasize the potential of HoloCNN-SHO as a reliable, accurate, and clinically applicable technique for efficient liver cancer detection and screening.