Skip to content

Releases: mohdsinad/FRMIS-Python

Version 1.1.0

Choose a tag to compare

@mohdsinad mohdsinad released this 24 Nov 08:43

Full Changelog: 1.0.1...1.1.0

Version 1.0.1

Choose a tag to compare

@mohdsinad mohdsinad released this 20 Nov 10:38

Full Changelog: 1.0...1.0.1

Version 1.0

Choose a tag to compare

@mohdsinad mohdsinad released this 04 Nov 06:39

Features

  • Robust Pairwise Alignment: Uses feature detection (SIFT) and RANSAC to accurately calculate translations between adjacent images.
  • Global Optimization: Employs a Minimum Spanning Tree (MST) approach on a graph of image connections to find the optimal global position for each tile, minimizing accumulated error.
  • Blending Options: Supports both simple overlay and weighted linear blending to create seamless transitions between images.
  • Command-Line Interface: Easy-to-use CLI for running the stitching process with configurable parameters.
  • Grid-Based: Optimized for image sets captured in a grid-like pattern (e.g., microscopy, aerial photography).

Usage

The tool is run from the command line. You must provide parameters specifying the input image directory, output path, and image grid dimensions.

Command-Line Arguments

Argument Shorthand Description Required
--dataset_dir -d Path to the directory containing the input images. Yes
--output_path -o Path to save the final stitched mosaic image. Yes
--mosaic_height -mh The number of rows in the image grid. Yes
--mosaic_width -mw The number of columns in the image grid. Yes
--image_height -ih The height of a single image in pixels. Yes
--image_width -iw The width of a single image in pixels. Yes
--overlap_x -ox The approximate horizontal overlap between adjacent images in pixels. Yes
--overlap_y -oy The approximate vertical overlap between adjacent images in pixels. Yes
--image_channels -ic The number of channels in the images (e.g., 3 for RGB). Default: 3. No
--blend -b The blending method to use (overlay or linear). Default: linear. No

Full Changelog: https://github.com/mohdsinad/FRMIS-Python/commits/1.0