Early Cancer Detection Combining Computer Artificial Intelligence with CT Medical Imaging: Application of Big Data Approach

Traditional imaging examinations, blood tests, and biomarker tests often have problems with insufficient sensitivity and specificity. To this end, this paper focuses on artificial intelligence, especially the application of Deep Learning (DL) models in CT (computed tomography) image analysis and diagnosis, and explores its potential in early cancer detection. In this study, Convolutional Neural Network (CNN) is applied to tumor detection in CT images. First, the image data is standardized, including resizing and normalization, to make sure that the input data is consistent, and more training samples are generated through methods for augmenting data (e.g., cropping, flipping, rotation, etc.) to enhance the model's capacity for generalization. The CNN model is primarily composed of convolutional layers, layers that pool together, fully linked layers, and functions for activation. The convolutional layers automatically extract low-level features (such as edges, textures, etc.) in the image through local receptive fields, and the pooling layers further reduce the feature dimension and enhance the model's anti-noise ability. Finally, the experimental data show that the method in this paper achieves an accuracy of 96.63 and a prediction time of 1.81 seconds, which is almost comparable to that of the professional doctor, but it is substantially shorter in prediction time.