Image Processing

Multimedia Processing Traning in Summer, 2019

---

Project List

• Image_pointer : Image Rotation with Pointer
• DCT : Discrete Cosine Transform and Reconstruct
• DFT : Discrete Fourier Transform and Reconstruct
• Interpolation-Sample_and_hold : upscaling, downscaling with Sample and hold interpolation
• Interpolation-Bilinear : upscaling, downscaling with Bilinear interpolation
• Interpolation-cubic_convolution : upscaling, downscaling with Cubic convolution interpolation
• Interpolation-cubic_B_Spline : upscaling, downscaling with Cubic B Spline interpolation

---

Getting Started

Stack

• C

Pre-requisties

Requirement	Description
visual studio	v2017 or higher
git	We follow the Github flow

Must to do

• you have to include raw image file in your working directory.
• I.E) if you work on dct-idct, you have to copy lena image in dct-idct\Project64

Docs

display

• this includes raw image display program
  • RDisp
  • yuvplayer : Size -> Custom, Color -> Y

img

• this is raw image folder which are the most frequently used in image processing, including lena, baboon, and F16

src

• img_pointer

  • You can get used to pointer by vertical mirroring, horizon mirroring, and rotating
    • arr.c : image transformation with array
    • main_pointer.c : image transformation with pointer

• dct-idct

  • You can understand lossy and loseless compression by practicing discrete cosine transform(DCT) and reconstruction(Inverse DCT - IDCT).
    • default : 8X8 dct block (block param : N in dct.h)
    • dct_quantization.c : reconstruct with 35dB, 58dB, 28dB, and 30dB quantization coefficient
    • idct_coefficient.c : reconstruct with input coefficient
    • dct_direction.c : dct test
    • psnr.c : you can check how similar with original image and reconstruct image add below code in main.c

      psnr(input, outimg);
      

      if (menu == 1) {
            DCT(input, output);
            print(output, outimg);
      
            FILE * outfile = fopen("lena_defaultDCT.img", "wb");
            fwrite(outimg, sizeof(char), HEIGHT*WIDTH, outfile);
            psnr(input, outimg);
      
        fclose(outfile);
            free(output);
            free(outimg);
      
            printf("success!");
        }
      

• dft-idft

  • You can understand Fourier Transform in Image Processing
    • DCT.c : Discrete Cosine Transform
    • DFT.c : Discrete Fourier Transform
    • Filter.c : LPF

• interpolation

  • You can choice 4 interpolation algorithm when upscaling, downscaling, or rotating images
  • You can upscale img, downscale img, upscale img and reconstruct, and downscale img and reconstruct with or without LPF
    • interpolation-sample_and_hold
    • interpolation-bilinear
      • filter.c : pre-calculate and declare filter
      • makeFilter.c : you can make LPF with this code
    • interpolation-cubic_convolution
    • interpolation-cubic_Bspline
      • main.c : basic main function
      • menu.c : updated main function with menu