https://github.com/flexrobotics/roboflex_util_jpeg

Roboflex in-memory compression of rgb images to and from jpegs.
https://github.com/flexrobotics/roboflex_util_jpeg

jpeg roboflex

Last synced: 2 months ago
JSON representation

Roboflex in-memory compression of rgb images to and from jpegs.

• Host: GitHub
• URL: https://github.com/flexrobotics/roboflex_util_jpeg
• Owner: flexrobotics
• License: mit
• Created: 2023-10-19T21:24:08.000Z (over 1 year ago)
• Default Branch: main
• Last Pushed: 2023-11-04T20:39:21.000Z (over 1 year ago)
• Last Synced: 2025-02-23T00:19:24.947Z (3 months ago)
• Topics: jpeg, roboflex
• Language: C++
• Homepage: https://github.com/flexrobotics/roboflex_util_jpeg
• Size: 15.6 KB
• Stars: 1
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# roboflex.util.jpeg

Roboflex support for jpeg compression to file and memory, and jpeg decompression from memory.

Useful for compressing images over slow transports (wifi, etc).

## System Dependencies

    None! We build jpeg-compression from source...

## pip install

    pip install roboflex.util.jpeg

## Import

    import roboflex.util.jpeg as ruj

## Nodes

There are two complementary nodes: `JPEGCompressor`, which can turn rgb tensors into jpegs in memory, and `JPEGDecompressor`, which does the opposite. Useful either for writing rgb tensors to files, or compressing them (NOTE jpeg is lossy).

    # all parameters are optional

    c = ruj.JPEGCompressor(

        # in the incoming message, where to find the rgb tensor

        image_key = "rgb",

        # in the outgoing message, where to place the jpeg data

        output_key = "jpeg", 

        # If this is provided, will ALSO write jpeg files with this 

        # prefix, with a variation of the date and time as the suffix.

        filename_prefix = "",

        # name of the node

        name = "JPEGCompressor",

        # prints internal info

        debug = False,

    )

... and ...

    c = ruj.JPEGDecompressor(

        # in the incoming message, where to find the jpeg data as a blob

        input_key = "jpeg",

        # in the outgoing message, where to place the rgb data as a tensor

        output_key = "rgb", 

        # name of the node

        name = "JPEGDecompressor",

        # prints internal info

        debug = False,

    )