Specialization_Capstone_Project - Image Classification

This project is designed to classify multiple TIFF images stored in a specified folder using the cuDNN library. The program starts by initializing the necessary CUDA and cuDNN resources and checks the system's GPU capabilities. It then iterates through all TIFF files in the given directory, loading and preprocessing each image into a suitable format for neural network inference. The image data is normalized and stored in a tensor descriptor. The core of the program involves a simplified placeholder for running a neural network model on the preprocessed image data to obtain classification results. These results, including predicted class labels and associated probabilities, are written to an output text file. The program ensures proper memory management by freeing allocated resources and destroys the cuDNN handle upon completion. This structure allows for batch processing of images with efficient use of GPU resources for deep learning tasks.