Deep Learning-Based Techniques for Cancer Cell Image Recognition

Abstract

Cancer remains one of the most frightening threats to human health. It is often difficult to detect in its early stages, and currently, there are almost no effective treatment methods, while medical costs remain extremely high. To address this problem, this paper proposes a “deep-learning-based cancer cell image recognition method.” By collecting a large number of images from the internet, we design and train deep neural networks (DNNs) and convolutional neural networks (CNNs), adjusting their parameters to evaluate whether these models can accurately distinguish between cancerous and normal cells. We used two datasets: one containing 2,000 images and another containing 4,000 images. Both models were trained separately on the two datasets. The results show that the DNN model achieved accuracies of 72% and 73%, respectively-an improvement of 1%. The CNN model achieved accuracies of 75% and 78%, corresponding to a 3% improvement. These findings demonstrate that CNNs are more effective than DNNs for cancer cell image classification, and that larger datasets significantly improve model accuracy.