digital-image-processing

Here are the programs that I wrote during the Digital Image Processing class in 2008, as well as some new experiments.

1. Grayscale to black-and-white
   1. Thresholding
   2. Bit-plane slicing
   3. Random dithering
   4. Ordered dithering
      1. Bayer dithering
      2. Clustered dithering
   5. Error-diffusion dithering
      1. Floyd-Steinberg dithering
      2. Jarvis-Judice-Ninke dithering
      3. Atkinson dithering
   6. Image-based dithering
      1. Halftoning with a single texture
      2. Halftoning with texture fusion
2. Image scaling
   1. Nearest-neighbor interpolation
   2. Bilinear interpolation
3. Alpha compositing
4. Intensity transformations
   1. Gamma correction
   2. Linear contrast stretching
   3. Histogram equalization
      1. Histogram equalization
      2. Adaptive histogram equalization
      3. Contrast-limited adaptive histogram equalization
5. Convolution
6. Median filter
7. Segmentation

---

Thresholding

Input	Output (threshold = 100)

---

Bit-plane slicing

Output (bit 6)	Output (bit 5)

Output (bit 4)	Output (bit 3)

---

Random dithering

Input	Output (threshold = 100, noise = 40)

---

Bayer dithering

Input	Output (4x4 Bayer matrix)

---

Clustered dithering

Input	Output

---

Floyd-Steinberg dithering

Input	Output

---

Jarvis-Judice-Ninke dithering

Input	Output

---

Atkinson dithering

Input	Output

---

Halftoning with a single texture

Input	Output (block size = 8, alpha = 0.6)

---

Halftoning with texture fusion

Input	Output (block size = 5, alpha = 0.7, threshold1 = 110, threshold2 = 140)

---

Nearest-neighbor interpolation

Input	Output (N = 5, cropped)

---

Bilinear interpolation

Input	Output (N = 5, cropped)

---

Alpha compositing

Input	Output
Goldhill.png as top Lena.png as bottom Peppers.png as alpha channel

---

Gamma correction

Input	Output (c = 1.0, gamma = 0.6)

Input	Output (c = 1.5, gamma = 1.5)

---

Linear contrast stretching

Input	Output (delta = 1000)

---

Histogram equalization

Input	Output

---

Adaptive histogram equalization

Input	Output (block size = 100)

---

Contrast-limited adaptive histogram equalization

Input	Output (block size = 64, clip limit = 2.2)

---

Convolution

Input	Output (11x11 blur kernel)

Input	Output (Sharp kernel)

Input	Output (Vertical Sobel kernel)

---

Median filter

Input	Output (3x3 plus aperture)

---

Segmentation

Input	Output (threshold = 128)
	12 segments

Input (cropped)	Output (threshold = 80, cropped)
	397 segments