Adaptive Quadtree Refinement and Matching Pursuit

Código para Adaptive Quadtree Refinement and Matching Pursuit (AQMP)

Installation

You can install the Image Compressor package via pip:

pip install aqmp-compressor

Usage

Command Line

You can run the image compressor from the command line using the following command:

python main.py name_of_image.png --min_sparcity=0.1 --min_n=8 --max_n=32 --a_cols=64 --max_error=0.01 --wavelet_election=db1 --shuffle_dictionary=False --v_format_precision=f --apply_deflate=False

If no parameters are chosen, these are the predefined ones:

• min_sparcity=0.1
• min_n=8
• max_n=32
• a_cols=64
• max_error=0.01
• wavelet_election=db1
• shuffle_dictionary=False
• v_format_precision=f
• apply_deflate=False

Importing the ImageCompressor Class

You can also use the ImageCompressor class directly in your Python code. Here is an example of how to use it in a Jupyter Notebook:

from aqmp import ImageCompressor

# Initialize the compressor with default parameters
compressor = ImageCompressor(
    min_sparcity=0.1,
    min_n=8,
    max_n=32,
    a_cols=64,
    max_error=0.01,
    wavelet_election="db1",
    shuffle_dictionary=False,
    v_format_precision="f",
    apply_deflate=False
)

# Compress the image
input_file = 'path/to/your/image.png'
output_file = 'path/to/save/compressed_image.fif'
compressor.encode(input_file, output_file)

# Decompress the image
input_file = 'path/to/save/compressed_image.fif'
output_file = 'path/to/save/decompressed_image.png'
compressor.decode(input_file, output_file)

This example demonstrates how to compress and decompress an image using the ImageCompressor class.

Roadmap:

• Compare with JPEG Algorithm
• Fix dicctionary product with subimages
• Print or show different compression levels
• Cambiar nombre .fif (Fast Image Format) por otro más sugerente del algoritmo actual
• Implement optuna optimization
• Check alternative dictionaries
  • Adaptive dictionaries: DCT or wavelet bases, K-means clustering-based dictionary.
  • Non-orthogonal basis functions: Gabor wavelets or curvelets.
• Test different compression levels with sparsity parameter.
• Implement DEFLATE function from zlib to see check if there are changes in the SSIM index.
• Extend to video compressive sensing (2nd paper?) [https://www.mdpi.com/2076-3417/12/5/2734]

Encoder Diagram:

Images:

Test images have been taken from Here. Filenames and description use are the following

4.1.01 Female (NTSC test image) 256 Color 4.1.02 Couple (NTSC test image) 256 Color 4.1.03 Female (from Bell Labs?) 256 Color 4.1.04 Female 256 Color 4.1.05 House 256 Color 4.1.06 Tree 256 Color 4.1.07 Jelly beans 256 Color 4.1.08 Jelly beans 256 Color 4.2.01 Splash 512 Color 4.2.03 Mandrill (a.k.a. Baboon) 512 Color 4.2.05 Airplane (F-16) 512 Color 4.2.06 Sailboat on lake 512 Color 4.2.07 Peppers 512 Color