Changelog

All notable changes to U-FISH will be documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[Unreleased]

Added

• Comprehensive Read the Docs documentation

• Support for multi-GPU training

• Batch processing optimizations

• New tutorial notebooks

Changed

• Improved memory efficiency for large images

• Enhanced 3D processing performance

Fixed

• Memory leak in tiling mode

• Edge case in spot matching algorithm

[1.0.0] - 2024-01-15

This is the first stable release of U-FISH.

Added

• Core FISH spot detection functionality

• U-Net based image enhancement

• Support for 2D and 3D images

• Pre-trained models for various FISH types

• Python API for programmatic access

• Command-line interface

• Evaluation metrics and visualization tools

• Support for large-scale data formats (Zarr, N5, OME-Zarr)

• GPU acceleration with ONNX Runtime

• Comprehensive test suite

• Example notebooks and tutorials

Features

• High accuracy: State-of-the-art F1 scores on benchmark datasets

• Small model: Only 160k parameters (680KB ONNX file)

• Fast inference: Real-time processing on GPU

• Flexible: Works with diverse FISH imaging protocols

• User-friendly: Simple API and CLI

[0.9.0] - 2023-12-01

Beta release for testing and feedback.

Added

• Training and fine-tuning capabilities

• Data augmentation pipeline

• Mixed precision training support

• Early stopping and learning rate scheduling

• Model checkpointing

• TensorBoard integration

Changed

• Improved model architecture for better accuracy

• Optimized data loading pipeline

• Enhanced CLI with more options

Fixed

• Numerical stability in loss calculation

• Memory issues with large batch sizes

[0.8.0] - 2023-10-15

Alpha release with core functionality.

Added

• Basic U-Net implementation

• Simple spot detection algorithm

• Initial API design

• Basic CLI commands

• Preliminary documentation

Changed

• Switched from TensorFlow to PyTorch

• Adopted ONNX for model deployment

[0.7.0] - 2023-08-01

Pre-alpha development version.

Added

• Project structure

• Initial U-Net experiments

• Dataset collection and curation

• Benchmark framework

Migration Guides

Migrating from 0.x to 1.0

The 1.0 release includes several API changes:

1. Model Loading:

   # Old (0.x)
   model = UFish.load_model('path/to/model')
   
   # New (1.0)
   ufish = UFish()
   ufish.load_weights('path/to/model')
   

2. Prediction API:

   # Old (0.x)
   spots = model.predict(image)
   
   # New (1.0)
   spots, enhanced = ufish.predict(image)
   

3. CLI Commands:

   # Old (0.x)
   ufish-predict input.tiff output.csv
   
   # New (1.0)
   ufish predict input.tiff output.csv
   

Deprecation Policy

• Features marked as deprecated will be maintained for at least 2 minor versions

• Deprecation warnings will be issued when using deprecated features

• Migration guides will be provided for all breaking changes

Future Releases

Planned features for future releases:

Version 1.1.0 (Q2 2024) * Improved 3D processing algorithms * Support for multi-channel prediction * Performance optimizations * Additional pre-trained models

Version 1.2.0 (Q3 2024) * Integration with popular image analysis platforms * Advanced visualization features * Distributed processing support * Model uncertainty estimation

Version 2.0.0 (2025) * Next-generation architecture * Transformer-based models * Self-supervised pre-training * Real-time video processing

Versioning Policy

U-FISH follows Semantic Versioning:

• MAJOR version for incompatible API changes

• MINOR version for backwards-compatible new features

• PATCH version for backwards-compatible bug fixes

Release Schedule

• Patch releases: As needed for bug fixes

• Minor releases: Quarterly

• Major releases: Annually

Support Policy

• Latest version: Full support

• Previous minor version: Security and critical bug fixes

• Older versions: Community support only

For the complete changelog and release notes, visit the GitHub Releases page.