# Imagcompmatlab is a preliminary draft that can be used to study image compression using the DCT transform. This script is designed for research purposes without considering computational overhead.

The paper can be found here
Available: https://cipn.ku.edu.np/?page_id=372

Cite the code as:
P. Devkota, “DCT-Image-compression.” Accessed: Jul. 01, 2023. [Online]. Available: https://github.com/Prabal1998/DCT-Image-compression



Run main.m file. 





A user shall input the required function argument as directed and hinted below.
A MATLAB termainal ip/op is  included herewith.



ans =

    'Make sure directory of image file and executing MATLAB script file is same: '

enter the file name, input string inside single quotation mark: 
=
    'marbles4.tiff'

x =

    'marbles4.tiff'


ans = 

  struct with fields:

                     Filename: '..\marbles4.tiff'
                  FileModDate: '03-Jan-2023 15:22:21'
                     FileSize: 1030656
                       Format: 'tif'
                FormatVersion: []
                        Width: 840
                       Height: 654
                     BitDepth: 24
                    ColorType: 'truecolor'
              FormatSignature: [73 73 42 0]
                    ByteOrder: 'little-endian'
               NewSubFileType: 0
                BitsPerSample: [8 8 8]
                  Compression: 'LZW'
    PhotometricInterpretation: 'RGB'
                 StripOffsets: [8 5848 11771 17729 23763 29953 36076 42098 48117 54112 60059 66118 72256 78510 84672 90840 … ]
              SamplesPerPixel: 3
                 RowsPerStrip: 4
              StripByteCounts: [5840 5923 5958 6034 6190 6123 6022 6019 5995 5947 6059 6138 6254 6162 6168 5998 6119 6190 … ]
                  XResolution: 96
                  YResolution: 96
               ResolutionUnit: 'Inch'
                     Colormap: []
          PlanarConfiguration: 'Chunky'
                    TileWidth: []
                   TileLength: []
                  TileOffsets: []
               TileByteCounts: []
                  Orientation: 1
                    FillOrder: 1
             GrayResponseUnit: 0.0100
               MaxSampleValue: [255 255 255]
               MinSampleValue: [0 0 0]
                 Thresholding: 1
                       Offset: 1029136
                    Predictor: 'Horizontal differencing'


ans = 

  struct with fields:

                     Filename: '..\marbles4.tiff'
                  FileModDate: '03-Jan-2023 15:22:21'
                     FileSize: 1030656
                       Format: 'tif'
                FormatVersion: []
                        Width: 840
                       Height: 654
                     BitDepth: 24
                    ColorType: 'truecolor'
              FormatSignature: [73 73 42 0]
                    ByteOrder: 'little-endian'
               NewSubFileType: 0
                BitsPerSample: [8 8 8]
                  Compression: 'LZW'
    PhotometricInterpretation: 'RGB'
                 StripOffsets: [8 5848 11771 17729 23763 29953 36076 42098 48117 54112 60059 66118 72256 78510 84672 90840 … ]
              SamplesPerPixel: 3
                 RowsPerStrip: 4
              StripByteCounts: [5840 5923 5958 6034 6190 6123 6022 6019 5995 5947 6059 6138 6254 6162 6168 5998 6119 6190 … ]
                  XResolution: 96
                  YResolution: 96
               ResolutionUnit: 'Inch'
                     Colormap: []
          PlanarConfiguration: 'Chunky'
                    TileWidth: []
                   TileLength: []
                  TileOffsets: []
               TileByteCounts: []
                  Orientation: 1
                    FillOrder: 1
             GrayResponseUnit: 0.0100
               MaxSampleValue: [255 255 255]
               MinSampleValue: [0 0 0]
                 Thresholding: 1
                       Offset: 1029136
                    Predictor: 'Horizontal differencing'

Select either of matrix from: R1,G1,B1,Y,U,V,Y2,Cb,Cr,Y1,I,Q,Gray1,Gray2:
=
    Gray1

enter block size N representing N*N pixel:
=
    8

select input image matrix existing in workspace to plot: 
=
    Gray1

enter the number of row to plot intensity value of pixel element consisting along rows:
=
    10

select dct matrix to plot: 
=
    mat2quant

Enter the scaling factor f that is  divisor to the DCT coefficient: 
=
    5

scal_f =

     5

Enter the offset value to be applied to DCT coefficient:
=
    min(min(mat2quant))

pos_off =

     -290.7736

enter the number of bits to represent decision level for llyod quantizer:
=
    8

select input image matrix existing in workspace to write input image in binary format 
 :
 =
    Gray1
 
Create the encoded image huffman code file of .bin extension inside single quotation:
=
    'opp.bin'
  
enter the encoded image huffman dict file name (.bin extension). You can use same huffman 
 code file to append the dictionary result : 
 =
    'opp.bin'
 
Create the input test image file  (.bin extension) inside single quotation 
=
    'ipp.bin'
  
enter the input image type uint8 or uint16 or uint24 bits per sample in single quotatioin:
=
    'uint8'

ipimg_btspersam =

    'uint8'

Select mat2inv matrix existing in workspace to perform inverse DCT/DWT: 
=
    mat2inv

enter block size N representing N*N pixel to perform inverse DCT/DWT: 
=
    8

Select either of matrix from: R1,G1,B1,Y,U,V,Y2,Cb,Cr,Y1,I,Q,Gray1,Gray2:  
=
    Gray1

enter the input image maximum intensity value: 
=
              255

    -----mean square error---
            51.7944

    ---Peak signal to noise ratio---
            30.9880

enter the input image file name inside single quotation: 
=
    'input'
enter the output image file name inside single quotation:
=
    'output'
>>