EE 569 Digital Image Processing

My homework submissions for the digital image processing class @ USC

Homework 1

1. Image Demosaicing (Bilinear Demosaicing) and Histogram Manipulation

2. Image Denoising

    + Linear Filters (Uniform, Gaussian)
    + Bilinear Filters

3. Special Effect Imgage Filters (Oil painting Effect)

---

Homework 2

1. Edge Detection (Canny, Sobel and Structured Edge)

2. Digital Half-toning

3. Half-toning with Error Diffusion

---

Homework 3

1. Geometric Image Modification

2. Homographic Image Transformation and Image Stitching

3. Morphological Processing (Not implemented)

---

Homework 4

1. Texture Analysis

    + Feature Extraction 
    + Advanced Texture Classification (Unsupervised : KMeans), (Supervised : Random Forest)

2. Texture Segmentation

3. SIFT and Image Matching (Using OpenCV)

    + Salient Point Descriptor
    + Image Matching
    + Bag of Words (using KMeans Classifier)

---

Homework 5

CNN Training using LeNET-5

     + MNIST DATASET
     + FASHION MNIST DATASET
     + CIFAR-10 DATASET

Homework 6

1. Feedforward designed Convolutional Neural Networks (FF-CNN)

2. PixelHop and PixelHop++ (variations in using Swaab and CSwaab transforms)

3. MNIST and FASHION-MNIST Classifications

     + Building the PixelHop
     + Comparison between PixelHop and PixelHop++
     + Error Analysis