Built a CNN-based model to classify seven facial emotions (Angry, Disgust, Fear, Happy, Neutral, Sad, Surprise) using a custom RGB/grayscale dataset with varying lighting. Used stratified sampling with an 80/10/10 split. Compared baseline, basic, and enhanced CNN models—enhanced CNN performed best due to deeper layers, batch normalization, dropout, and early stopping.
Seam carving is an advanced algorithm for content-aware image resizing, This project presents a comparison of three algorithms for Seam Carving: Brute Force, Greedy, and Dynamic Programming. Each algorithm removes low-energy vertical seams to resize images while preserving important visual content. The performance, time complexity, and output quality of the three approaches were analyzed and compared.
Brute Force Seam Carving Output, Image resized with key visual features preserved
