ImageJ Fiji Auto-Stitcher This ImageJ macro, designed for use within Fiji (https://fiji.sc/), automates the stitching of multiple images into a single large image. It's particularly useful for images acquired using a microscope with a tiling function. Important: This script requires Fiji, which is a distribution of ImageJ that includes many useful plugins, including the "Grid/Collection stitching" plugin used by this script. Plain ImageJ might not have all the required plugins. Download Fiji from https://fiji.sc/. ⚠️ Terms of Use and Citation By using this script, you agree to cite it in any resulting publications, presentations, or other academic work. If you use this script in your research and publish the results, please cite it as follows: Wiegand, M. (2025). ImageJ Fiji Auto-Stitcher (Version 3.0) [Computer software]. GitHub. https://github.com/wiegandlmu/ImageJ-Fiji-Auto-Stitcher Zenodo. https://doi.org/10.5281/zenodo.17401384 Recommended citation style for BibTeX: @software{wiegand2025autostitcher, author = {Wiegand, Max}, title = {ImageJ Fiji Auto-Stitcher}, year = {2025}, version = {3.0}, url = {https://github.com/wiegandlmu/ImageJ-Fiji-Auto-Stitcher}, note = {Computer software} } When you run the script for the first time, you will be asked to acknowledge this citation requirement. Citing this work allows others to find and utilize this tool and acknowledges the effort put into its development. Features NEW: Intelligent Pattern Recognition: Automatically detects filename patterns and numbering schemes without manual configuration NEW: Flexible File Naming: Handles different prefixes in different folders (e.g., Image_XY01_00001.tif vs H2_Image_XY01_00001.tif) NEW: Automatic Padding Detection: Recognizes different zero-padding lengths (00001, 0001, 001, etc.) Optimized Overlap: Uses a standard 20% overlap value that works well for most microscopy applications Optimized for Single Images: Streamlined for stitching individual 2D images (no Z-stacks) Sharp Stitching: Uses subpixel accuracy and optimized computation for precise overlap, resulting in sharp stitching results RGB Color Preservation: Automatically maintains color information in stitched images User-Friendly: Minimal configuration required - the script does the heavy lifting Automatic Grid Size Calculation: Intelligently calculates appropriate grid sizes based on image count What's New in Version 3.0? Version 3.0 introduces intelligent pattern recognition that eliminates the need for manual filename pattern configuration: Automatic Pattern Detection: The script now analyzes your files and automatically detects: File name structure and prefixes Which number sequence represents the tile numbers Zero-padding length (e.g., 00001 vs 0001) Handles Varying Prefixes: Works seamlessly even when different folders have different filename prefixes: Folder XY01: Image_XY01_00001_CH4.tif Folder XY02: H2_Image_XY02_00001_CH4.tif Folder XY03: Sample1_Image_XY03_00001_CH4.tif Enhanced Diagnostics: Provides detailed debug output showing what was detected Prerequisites Fiji: Download and install Fiji from https://fiji.sc/ How to Use 1. Acquire Images with Your Microscope Storage Format: Save images as TIFF (uncompressed or LZW-compressed) Tiling: Use the tiling function of your microscope to capture multiple images for stitching Folder Structure: The script expects images to be organized in subfolders within a main directory. Each subfolder should contain images belonging to one stitched image Example folder structure: MainFolder/ ├── XY01/ │ ├── Image_XY01_00001_CH4.tif │ ├── Image_XY01_00002_CH4.tif │ └── Image_XY01_00003_CH4.tif ├── XY02/ │ ├── H2_Image_XY02_00001_CH4.tif │ ├── H2_Image_XY02_00002_CH4.tif │ └── H2_Image_XY02_00003_CH4.tif └── XY03/ ├── Sample1_Image_XY03_00001_CH4.tif ├── Sample1_Image_XY03_00002_CH4.tif └── Sample1_Image_XY03_00003_CH4.tif The script will automatically detect the pattern in each folder! File Naming: Sequential numbering required: Files must have sequential numbers (1, 2, 3... or 00001, 00002, 00003...) Any prefix allowed: The script automatically detects prefixes Consistent numbering within folders: Each folder should use the same numbering pattern Different prefixes between folders: Each folder can have its own unique prefix Overlap: The script uses a standard 20% overlap value This works well for most microscopy applications Ensure your microscope is configured to capture images with approximately 15-25% overlap What is overlap? The percentage of the image area that overlaps with the neighboring image Example: With 20% overlap, a 1000×1000 pixel image overlaps 200 pixels with its neighbor If your microscope uses a significantly different overlap (e.g., 10% or 40%), you may need to modify the script (see "Advanced Settings" section) No Z-Stacks: This script is designed for single images only, not Z-stacks 2. Install and Run the Script in Fiji Download the script: If you are on the GitHub repository page, scroll up to the top Click the green "Code" button Select "Download ZIP" Save the ZIP file to your computer and unzip it Open the script in Fiji: Open Fiji Go to "Plugins" → "Macros" → "Run..." Navigate to the unzipped folder and select the AutoStitch_V3.0.ijm file Click "Open" Accept Citation Terms: A dialog will appear asking you to acknowledge the citation requirement Click "I agree to cite this script in my work" checkbox Click "OK" to proceed Configuration Dialog (Simplified!): A dialog box will appear with minimal configuration needed: File extension: The extension of your image files (default: ".tif") Grid Size: Specify the grid size (e.g., "3x3") or use "0x0" for automatic calculation Automatic calculation (0x0): The script calculates a square grid based on the number of images Example: 9 images → 3×3 grid, 16 images → 4×4 grid, 10 images → 4×4 grid (with 6 empty positions) ⚠️ Important: Automatic calculation assumes a square or nearly-square grid arrangement If your images are arranged in a rectangular grid (e.g., 2×6), you must enter the grid size manually Manual entry: Recommended if you know your exact grid dimensions (e.g., "2x6", "3x4") Click "OK" Select Main Folder: A new window will appear Navigate to the main folder and select it (the folder containing the subfolders with the images) Monitor Progress: The script will process each subfolder automatically Watch the Log window for detailed information about pattern detection and progress The log will show: Detected filename patterns for each folder Number of sequential images found Grid size being used Success/error messages Finding the stitched images: Stitched images will be saved in a new folder called Autostitch in your main directory Each stitched image is named after its source folder (e.g., XY01_stitched.tif) The final stitched image will also be displayed in Fiji Understanding the Pattern Detection The script automatically analyzes your files and detects the sequential numbering pattern. The actual filename doesn't matter - only the sequential numbers! Example 1: Simplest Possible Case Files: 1.tif, 2.tif, 3.tif, 4.tif Detected: Single number sequence Pattern: {i}.tif (1-digit) Works perfectly! Example 2: With Padding Files: 001.tif, 002.tif, 003.tif Detected: Zero-padded sequence Pattern: {iii}.tif (3-digit padding) Example 3: Standard Pattern Files: Image_XY01_00001_CH4.tif, Image_XY01_00002_CH4.tif Detected: Multiple number sequences found (XY01 and 00001) Identified: The sequence that increments (00001 → 00002) is the tile number Pattern: Image_XY01_{iiiii}_CH4.tif (5-digit padding) Example 4: Complex Custom Naming Files: MyExperiment_Sample3_Tile_00001_488nm.tif, MyExperiment_Sample3_Tile_00002_488nm.tif Detected: The incrementing sequence is "00001" → "00002" Pattern: MyExperiment_Sample3_Tile_{iiiii}_488nm.tif Everything else is just text - the script handles it automatically! Example 5: Different Folders, Different Filenames Folder XY01: Image_XY01_00001.tif Folder XY02: Scan2_Image_XY02_00001.tif Folder XY03: 1.tif, 2.tif, 3.tif Each folder is analyzed independently Different patterns are applied to each folder automatically The script doesn't care about folder names or filename prefixes - it just finds the sequential numbers! Troubleshooting "Could not detect valid filename pattern" message This usually means the script couldn't identify sequential numbering. Check: Sequential numbering: Files must be numbered 1, 2, 3... (or 00001, 00002, 00003...) starting from 1 No gaps: The sequence should be continuous (1, 2, 3, not 1, 3, 5) Consistent pattern: All files in a folder should use the same numbering and padding File extension: Ensure the file extension matches what you specified (default: .tif) The filename itself can be anything - only the sequential numbers matter! Example of good numbering: Good: 1.tif, 2.tif, 3.tif, 4.tif Good: Image_001.tif, Image_002.tif, Image_003.tif Good: Tile_00001.tif, Tile_00002.tif, Tile_00003.tif Good: abc_xyz_123_00001.tif, abc_xyz_123_00002.tif Example of problematic numbering: Bad: Image_1.tif, Image_3.tif, Image_7.tif (gaps in sequence) Bad: Image_0.tif, Image_1.tif, Image_2.tif (starts from 0 instead of 1) Bad: Image_a.tif, Image_b.tif (letters instead of numbers) Bad: Image_001.tif, Image_02.tif, Image_003.tif (inconsistent padding) "Too few sequential images" message The script found files but couldn't identify enough sequential images Check the log window for DEBUG information showing what was detected Ensure files are numbered starting from 1 (not 0) Verify zero-padding is consistent (all 00001 or all 0001, not mixed) Double or blurred edges in the stitched image The script uses a standard 20% overlap value that works for most applications If your microscope uses a significantly different overlap: Check your microscope settings to confirm the actual overlap percentage Modify the tileOverlap variable in the script (see "Advanced Settings" section below) Try values between 15-25% Black and white images instead of color This should not happen with version 2.0+ If it does, the script automatically converts to RGB before saving Out of memory errors The script uses the "Save computation time (but use more RAM)" option for best results If you have limited RAM, you can modify the script to use "Save memory (but be slower)" Advanced Settings (for advanced users) If needed, you can directly adjust the script settings. To make changes to the script: Open the AutoStitch.ijm file in a text editor (e.g., TextEdit on Mac, Notepad++ on Windows) Make the desired changes Save the file Run the modified script in Fiji Available parameters: Overlap Setting By default, the script uses 20% overlap. To change this, find this line in the script: // Fixed overlap at 20% (standard value for most applications) tileOverlap = "20"; Change "20" to your desired overlap percentage (e.g., "15" or "25"). Disabling Automatic Pattern Detection If you want to use a manual pattern (like in older versions), you can modify the script. However, the automatic detection is recommended for most users. Other Parameters These are usually fine and should not be changed unless needed: fusionMethod: Method for blending overlapping regions (default: "[Linear Blending]") regThreshold: Regression threshold for the stitching algorithm (default: 0.30) maxAvgThreshold: Maximum/average displacement threshold (default: 2.50) absThreshold: Absolute displacement threshold (default: 3.50) computation_parameters: Set to "[Save computation time (but use more RAM)]" for best results imageOutput: How the output should be handled (default: "[Fuse and display]") Find further information on how to adjust the script here: https://imagej.net/plugins/image-stitching Changelog Version 3.0 (2025) NEW: Intelligent Pattern Recognition: Automatically detects filename patterns NEW: Flexible Prefix Handling: Works with different prefixes in different folders NEW: Automatic Padding Detection: Recognizes zero-padding length automatically NEW: Ultimate Filename Flexibility: Works with any filename as long as sequential numbering is present (even simple 1.tif, 2.tif, 3.tif) Eliminated need for manual filename pattern configuration Enhanced diagnostic output with detailed pattern detection information Improved robustness for diverse naming conventions Output files now named after source folders for better organization Complete rewrite of pattern detection algorithm for maximum flexibility Version 2.1 (2025) Simplified user interface by removing overlap setting from dialog Set standard overlap to 20% (optimal for most microscopy applications) Added advanced settings section in README for custom overlap values Improved citation format with BibTeX support Version 2.0 (2025) Fixed blurry stitching results by optimizing computation parameters Added subpixel accuracy for sharper tile transitions Fixed black and white output by adding RGB conversion Added automatic overlap computation Improved user interface with flexible configuration dialog Added citation requirement dialog Simplified code and removed unnecessary functions Better console output for debugging License This project is licensed under the MIT License - see the LICENSE file for details. Acknowledgments This script uses the "Grid/Collection stitching" plugin in Fiji which is based on the publication: Preibisch, S., Saalfeld, S., & Tomancak, P. (2009). Globally optimal stitching of tiled 3D microscopic image acquisitions. Bioinformatics, 25(11), 1463–1465. https://imagej.net/plugins/image-stitching Thanks to all contributors of the ImageJ and Fiji community Remember: If you use this script in your research, please cite it! Your citation helps others discover this tool and supports continued development.