DeepTrack2 2.0.0

Stable release of DeepTrack2 2.0.0
Release date: 26 August 2024

PyPi 2.0.0
https://pypi.org/project/deeptrack/2.0.0/

This major release is a significant update that includes the migration to PyTorch and integration with the deeplay library. This version focuses on PyTorch/Deeplay integration, performance optimization, and a more rigorous approach to utilizing static data from disk.

Breaking Changes

Image Objects

• Pipelines no longer return Image objects by default, to improve performance and compatibility with other libraries. This change will not affect you unless you directly interact with the .properties attribute or call .get_properties(.) on pipeline outputs.
• If you need Image objects, you can enable them by using pipeline.store_properties(), restoring the behavior from previous releases.

TensorFlow Dependency Removed

• TensorFlow has been removed as a dependency due to compatibility issues and the complexity of maintaining it alongside major Python versions, especially on Windows.
• We will not actively support TensorFlow versions newer than 2.10. Users are encouraged to manage TensorFlow installations independently.

What's New

Global Changes

• Many submodules are now lazily loaded, only initializing when needed. This change significantly speeds up initial load times and reduces TensorFlow-related issues for users not requiring neural network functionalities.

Sources

• Introducing Sources, a new method for efficiently operating on static datasets. This feature addresses common problems with data manipulation and pipeline evaluation, allowing for more functional and efficient dataset handling.

PyTorch Integration

• PyTorch code is available in the pytorch submodule, which is lazy-loaded to prevent unnecessary overhead. This update includes:
  • pytorch.Dataset: A subclass of torch.utils.data.Dataset, making DeepTrack pipelines compatible with standard DataLoaders.
  • pytorch.ToTensor: A feature for converting pipeline outputs to PyTorch tensors, with support for specifying the dtype.

Performance Optimization

• This release is centered on performance optimization and a more effective way of using static data, enhancing the user experience and streamlining workflows.

New Optics

• Added iscat, darkfield, and improved holography imaging modalities.

deeplay

• The deeplay library is now accessible as deeptrack.deeplay, offering a powerful and flexible foundation for constructing neural networks with PyTorch.

This update marks a significant milestone in DeepTrack's development, emphasizing our commitment to performance, usability, and modern deep learning practices. We're excited to see how these improvements will enable our users to achieve even more with DeepTrack.

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891

DeepTrack2 2.0.0rc0

Stable release of DeepTrack2 2.0.0rc0
Release date: 19 March 2024

PyPi 2.0.0rc0
https://pypi.org/project/deeptrack/2.0.0rc0/

This major release is a significant update that includes the migration to PyTorch and integration with the deeplay library. This version focuses on PyTorch/Deeplay integration, performance optimization, and a more rigorous approach to utilizing static data from disk.

Breaking Changes

Image Objects

• Pipelines no longer return Image objects by default, to improve performance and compatibility with other libraries. This change will not affect you unless you directly interact with the .properties attribute or call .get_properties(.) on pipeline outputs.
• If you need Image objects, you can enable them by using pipeline.store_properties(), restoring the behavior from previous releases.

TensorFlow Dependency Removed

• TensorFlow has been removed as a dependency due to compatibility issues and the complexity of maintaining it alongside major Python versions, especially on Windows.
• We will not actively support TensorFlow versions newer than 2.10. Users are encouraged to manage TensorFlow installations independently.

What's New

Global Changes

• Many submodules are now lazily loaded, only initializing when needed. This change significantly speeds up initial load times and reduces TensorFlow-related issues for users not requiring neural network functionalities.

Sources

• Introducing Sources, a new method for efficiently operating on static datasets. This feature addresses common problems with data manipulation and pipeline evaluation, allowing for more functional and efficient dataset handling.

PyTorch Integration

• PyTorch code is available in the pytorch submodule, which is lazy-loaded to prevent unnecessary overhead. This update includes:
  • pytorch.Dataset: A subclass of torch.utils.data.Dataset, making DeepTrack pipelines compatible with standard DataLoaders.
  • pytorch.ToTensor: A feature for converting pipeline outputs to PyTorch tensors, with support for specifying the dtype.

Performance Optimization

• This release is centered on performance optimization and a more effective way of using static data, enhancing the user experience and streamlining workflows.

New Optics

• Added iscat, darkfield, and improved holography imaging modalities.

deeplay

• The deeplay library is now accessible as deeptrack.deeplay, offering a powerful and flexible foundation for constructing neural networks with PyTorch.

This update marks a significant milestone in DeepTrack's development, emphasizing our commitment to performance, usability, and modern deep learning practices. We're excited to see how these improvements will enable our users to achieve even more with DeepTrack.

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891

DeepTrack2 1.7.0 (LEGACY RELEASE)

Stable release of DeepTrack2 1.7.0 (LEGACY RELEASE)
Release date: 19 March 2024

PyPi 1.7.0
https://pypi.org/project/deeptrack/1.7.0/

This is a LEGACY RELEASE of DeepTrack2, which will be temporarily maintained to provide support and ensure compatibility for existing projects using TensorFlow. This release shall serve as a bridge for users transitioning to newer versions, enabling them to update their workflows at their own pace while still having access to critical fixes and updates.

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891

DeepTrack2 1.6.0

Stable release of DeepTrack2 1.6.0
Release date: 13 September 2023

PyPi 1.6.0
https://pypi.org/project/deeptrack/1.6.0/

DeepTrack2 1.6.0 is a minor release that mainly adds WAE (both WAE-GAN and WAE-MMD) and provides some minor edits to the code of GAN and VAEs. Additionally, it mitigates the impact of np.int deprecation on SampleToMasks and some other features.

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891

DeepTrack2 1.5.6

Stable release of DeepTrack2 1.5.6
Release date: 13 April 2023

PyPi 1.5.6
https://pypi.org/project/deeptrack/1.5.6/

DeepTrack2 1.5.6 is a patch that fixed an issue with the endothelial_vs dataset and an issue with repeated oneof-features.

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891

DeepTrack2 1.5.5

Stable release of DeepTrack2 1.5.5
Release date: 7 March 2023

PyPi 1.5.5
https://pypi.org/project/deeptrack/1.5.5/

DeepTrack2 1.5.5 is a patch

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891

DeepTrack2 1.5.4

Stable release of DeepTrack2 1.5.4
Release date: 18 February 2023

PyPi 1.5.4
https://pypi.org/project/deeptrack/1.5.4/

DeepTrack2 1.5.4 is a patch:
Fixed the reparameterization of VAEs to ensure proper computation.

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891

DeepTrack2 1.5.3

Stable release of DeepTrack2 1.5.3
Release date: 21 December 2022

PyPi 1.5.3
https://pypi.org/project/deeptrack/1.5.3/

DeepTrack2 1.5.3 is a patch:
Fixed an issue where ContinuousGraphGenerator yields an error if no detection is registered across multiple frames.

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891

DeepTrack2 1.5.2

Stable release of DeepTrack2 1.5.2
Release date: 13 December 2022

PyPi 1.5.2
https://pypi.org/project/deeptrack/1.5.2/

DeepTrack2 1.5.2 is a patch:

• Several improvements to the code of the attention layers.
• New datasets to train GNNs models and virtual staining neural networks.
• Added ability to approximate limited coherence length in Mie-optics.

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891

DeepTrack2 1.5.1

Stable release of DeepTrack2 1.5.1
Release date: 21 November 2022

PyPi 1.5.1
https://pypi.org/project/deeptrack/1.5.1/

DeepTrack2 1.5.1 is a patch:

• Fixed an issue using MieScatterer with the Repeat feature without wrapping it in Optics.
• Fixed an issue where MieSphere would get an incorrect phase for some optical parameters.
• Added a coherence_length parameter to MieScatterers to simulate limited spatial coherence.

Cite as:
Benjamin Midtvedt, Saga Helgadottir, Aykut Argun, Jesús Pineda, Daniel Midtvedt, Giovanni Volpe.
"Quantitative Digital Microscopy with Deep Learning."
Applied Physics Reviews 8, 011310 (2021)
https://doi.org/10.1063/5.0034891