Project information

• Category: Algorithm Design
• Location: Champaign, IL
• Project date: Nov 2023
• Project URL: temporarily unavailable

Reading features from images to automatically sort them into categories

• Developed and enhanced a neural network in PyTorch for classifying images into four categories by integrating convolutional neural network layers for spatial feature extraction in image data and optimizing with varied activation functions, and network architecture adjustments.
• Implemented cross-entropy loss function for multi-class classification and utilized PyTorch's automatic differentiation for efficient backpropagation.
• Ensured robust model generalization through L2 regularization and parameter tuning, while maintaining computational efficiency within a 500,000 value parameter constraint.