yansun1996/COGS260-Image-Recognition

Code for UCSD Cognitive Science 260 Image Recognition

COGS260-Image-Recognition

Code for UCSD CogSci 260 Image Recognition

Assignment 1

• OpenCV Basic (I/O, Smoothing, Denoising, Enhancement).

• Canny, Sobel and Structure Forest Edge Detection.

• Structure Forest Ref: https://github.com/ArtanisCV/StructuredForests.

Assignment 2

Assignment2 is finished in Python2.

• CNN, K-Nearest Neighbors (KNN) and Support Vector Machines (SVM) for MNIST classification.

• Spatial Pyramid Matching (SPM) + SVM for MNIST classification.

• SPM Ref: https://github.com/CyrusChiu/Image-recognition.

Assignment 3

• Least Square Estimation vs Parabola Estimation and L1-norm vs L2-norm.
• Perceptron Learning for Iris Dataset and compare the performance with raw data and Z-Score data.
• Feed Forward Neural Network for self-defined number of layers with self-defined number of neurons on MNIST.
• Modified AlexNet with Mini-batch training, evaluating different optimizers and accelerating tricks on CIFAR-10.

Assignment 4

• CharRNN for Shakespeare style text generation
• YOLO object detection with rotation and collages.