DeepImageJ-3.0.3

Release of DeepImageJ version 3.0.3.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

Changes with respect to DeepImageJ 2:

• Compatibility with Pytorch 2
• Compatibility with Tensorflow 2
• Compatibility with Onnx
• Changed backend to JDLL
• Improve inference speed and efficiency using ImgLib2

Changes with respect to DeepImageJ 3.0.1:

• Added links to relevant papers
• Correct engine location installation

Content of DeepImageJ_dependencies_3.zip:

• jna-5.12.0.jar
• dl-modelrunner-0.3.10.jar
• gson-2.10.1.jar
• imglib2-6.1.0.jar
• snakeyaml-1.33.jar

DeepImageJ-3.0.1

Release of DeepImageJ version 3.0.1.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

Changes with respect to DeepImageJ 2:

• Compatibility with Pytorch 2
• Compatibility with Tensorflow 2
• Compatibility with Onnx
• Changed backend to JDLL
• Improve inference speed and efficiency using ImgLib2

Content of DeepImageJ_dependencies_3.zip:

• jna-5.12.0.jar
• dl-modelrunner-0.3.10.jar
• gson-2.10.1.jar
• imglib2-6.1.0.jar
• snakeyaml-1.33.jar

DeepImageJ-2.1.16

Release of DeepImageJ version 2.1.16.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

Changes with respect to 2.1.15:

• Add compatibility to newer versions of Pytorch 1 (1.13, 1.12, 1.11, 1.10)

IMPORTANT (only for ImageJ1):

In order to run the plugin in GPU mode, the dependency libtensorflow_jni-1.15.0 has to be substituted by libtensorflow_jni_gpu-1.15.0. There should only be one of those two dependencies installed in the ImageJ1 directory.
DeepImageJ for GPU processing works even if there are is not any GPU available. However, the GPU version DOES NOT WORK on macOS distributions.

Content of DeepImageJ_dependencies_2116.zip:

• libtensorflow-1.15.0.jar
• libtensorflow_jni-1.15.0.jar
• proto-1.15.0.jar
• protobuf-java-3.5.1.jar
• junit-4.12.jar
• hamcrest-core-1.3.jar
• snakeyaml-1.21.jar
• slf4j-simple-1.7.26.jar
• slf4j-api-1.7.25.jar
• pytorch-jni-1.13.1-0.21.0.jar
• pytorch-engine-0.21.0.jar
• npy-0.3.3.jar
• kotlin-stdlib-1.3.72.jar
• jna-5.3.0.jar
• imagej-tensorflow-1.1.6.jar
• gson-2.8.5.jar
• commons-compress-1.20.jar
• api-0.21.0.jar

DeepImageJ-2.1.15

Release of DeepImageJ version 2.1.15.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

Changes with respect to 2.1.14:

• Increase robustness and correct bugs finding Pytorch torchscript and Tensorflow models in DIJ Run
• Update DIJ Install Model to be able to install all compatible models from the BioImage.io, previously only models produced by DeepImageJ could be installed.
• Fixed bug downloading models that contain forbidden characters on its name

IMPORTANT (only for ImageJ1):

In order to run the plugin in GPU mode, the dependency libtensorflow_jni-1.15.0 has to be substituted by libtensorflow_jni_gpu-1.15.0. There should only be one of those two dependencies installed in the ImageJ1 directory.
DeepImageJ for GPU processing works even if there are is not any GPU available. However, the GPU version DOES NOT WORK on macOS distributions.

Content of DeepImageJ_dependencies_2115.zip:

• libtensorflow-1.15.0.jar
• libtensorflow_jni-1.15.0.jar
• proto-1.15.0.jar
• protobuf-java-3.5.1.jar
• junit-4.12.jar
• hamcrest-core-1.3.jar
• snakeyaml-1.21.jar
• slf4j-simple-1.7.26.jar
• slf4j-api-1.7.25.jar
• pytorch-native-auto-1.9.1.jar
• pytorch-engine-0.14.0.jar
• npy-0.3.3.jar
• kotlin-stdlib-1.3.72.jar
• jna-5.3.0.jar
• imagej-tensorflow-1.1.6.jar
• gson-2.8.5.jar
• commons-compress-1.20.jar
• api-0.14.0.jar

DeepImageJ-2.1.14

Release of DeepImageJ version 2.1.14.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

The main changes with respect to DeepImageJ 1.2.1 are:

• From Image2image in 2D to image2any: deepImageJ 2.1.10 runs models defined for 3D image processing, image classification or object detection through region proposal networks. It deals with miscellaneous output formats and shapes.
• GPU support.
• Compatibility with PyTorch: All PyTorch models stored as TorchScript can be for the first time loaded into ImageJ/Fiji through deepImageJ.
• New formats for image pre and post-processing: ImageJ macros (.txt/.ijm) and Java code&classes (.jar/.class) are supported and can be combined.
• Compatibility with the Bioimage Model Zoo🦒: deepImageJ is a consumer of the trained models in the BioImage Model Zoo
• Addition of a Validation plugin
• Support of headless mode for ImageJ/Fiji
• Full support of PyImageJ
• Add of test feature in DIJ Run

No changes with respect to vesion 2.1.13. This version was published to tackle a conflict within the DeepImageJ ImageJ update site.

IMPORTANT (only for ImageJ1):

In order to run the plugin in GPU mode, the dependency libtensorflow_jni-1.15.0 has to be substituted by libtensorflow_jni_gpu-1.15.0. There should only be one of those two dependencies installed in the ImageJ1 directory.
DeepImageJ for GPU processing works even if there are is not any GPU available. However, the GPU version DOES NOT WORK on macOS distributions.

Content of DeepImageJ_dependencies_2110.zip:

• libtensorflow-1.15.0.jar
• libtensorflow_jni-1.15.0.jar
• proto-1.15.0.jar
• protobuf-java-3.5.1.jar
• junit-4.12.jar
• hamcrest-core-1.3.jar
• snakeyaml-1.21.jar
• slf4j-simple-1.7.26.jar
• slf4j-api-1.7.25.jar
• pytorch-native-auto-1.9.1.jar
• pytorch-engine-0.14.0.jar
• npy-0.3.3.jar
• kotlin-stdlib-1.3.72.jar
• jna-5.3.0.jar
• imagej-tensorflow-1.1.6.jar
• gson-2.8.5.jar
• commons-compress-1.20.jar
• api-0.14.0.jar

DeepImageJ-2.1.13

Release of DeepImageJ version 2.1.13.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

The main changes with respect to DeepImageJ 1.2.1 are:

• From Image2image in 2D to image2any: deepImageJ 2.1.10 runs models defined for 3D image processing, image classification or object detection through region proposal networks. It deals with miscellaneous output formats and shapes.
• GPU support.
• Compatibility with PyTorch: All PyTorch models stored as TorchScript can be for the first time loaded into ImageJ/Fiji through deepImageJ.
• New formats for image pre and post-processing: ImageJ macros (.txt/.ijm) and Java code&classes (.jar/.class) are supported and can be combined.
• Compatibility with the Bioimage Model Zoo🦒: deepImageJ is a consumer of the trained models in the BioImage Model Zoo
• Addition of a Validation plugin
• Support of headless mode for ImageJ/Fiji
• Full support of PyImageJ
• Add of test feature in DIJ Run

Changes with respect to 2.1.12:

• Adapt to changes in the rdf.yaml BioImage.io specs file
• Change the DJL API version to make them compatible with DeepImageJ again

IMPORTANT (only for ImageJ1):

In order to run the plugin in GPU mode, the dependency libtensorflow_jni-1.15.0 has to be substituted by libtensorflow_jni_gpu-1.15.0. There should only be one of those two dependencies installed in the ImageJ1 directory.
DeepImageJ for GPU processing works even if there are is not any GPU available. However, the GPU version DOES NOT WORK on macOS distributions.

Content of DeepImageJ_dependencies_2110.zip:

• libtensorflow-1.15.0.jar
• libtensorflow_jni-1.15.0.jar
• proto-1.15.0.jar
• protobuf-java-3.5.1.jar
• junit-4.12.jar
• hamcrest-core-1.3.jar
• snakeyaml-1.21.jar
• slf4j-simple-1.7.26.jar
• slf4j-api-1.7.25.jar
• pytorch-native-auto-1.9.1.jar
• pytorch-engine-0.14.0.jar
• npy-0.3.3.jar
• kotlin-stdlib-1.3.72.jar
• jna-5.3.0.jar
• imagej-tensorflow-1.1.6.jar
• gson-2.8.5.jar
• commons-compress-1.20.jar
• api-0.14.0.jar

DeepImageJ-2.1.12

Release of DeepImageJ version 2.1.12.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

The main changes with respect to DeepImageJ 1.2.1 are:

• From Image2image in 2D to image2any: deepImageJ 2.1.10 runs models defined for 3D image processing, image classification or object detection through region proposal networks. It deals with miscellaneous output formats and shapes.
• GPU support.
• Compatibility with PyTorch: All PyTorch models stored as TorchScript can be for the first time loaded into ImageJ/Fiji through deepImageJ.
• New formats for image pre and post-processing: ImageJ macros (.txt/.ijm) and Java code&classes (.jar/.class) are supported and can be combined.
• Compatibility with the Bioimage Model Zoo🦒: deepImageJ is a consumer of the trained models in the BioImage Model Zoo
• Addition of a Validation plugin
• Support of headless mode for ImageJ/Fiji
• Full support of PyImageJ
• Add of test feature in DIJ Run

Changes with respect to 2.1.11:

• Correct errors occurring with the "Test model" functionality

IMPORTANT (only for ImageJ1):

In order to run the plugin in GPU mode, the dependency libtensorflow_jni-1.15.0 has to be substituted by libtensorflow_jni_gpu-1.15.0. There should only be one of those two dependencies installed in the ImageJ1 directory.
DeepImageJ for GPU processing works even if there are is not any GPU available. However, the GPU version DOES NOT WORK on macOS distributions.

Content of DeepImageJ_dependencies_2110.zip:

• libtensorflow-1.15.0.jar
• libtensorflow_jni-1.15.0.jar
• proto-1.15.0.jar
• protobuf-java-3.5.1.jar
• junit-4.12.jar
• hamcrest-core-1.3.jar
• snakeyaml-1.21.jar
• slf4j-simple-1.7.26.jar
• slf4j-api-1.7.25.jar
• pytorch-native-auto-1.9.1.jar
• pytorch-engine-0.15.0.jar
• npy-0.3.3.jar
• kotlin-stdlib-1.3.72.jar
• jna-5.3.0.jar
• imagej-tensorflow-1.1.6.jar
• gson-2.8.5.jar
• commons-compress-1.20.jar
• api-0.15.0.jar

DeepImageJ-2.1.11

Release of DeepImageJ version 2.1.11.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

The main changes with respect to DeepImageJ 1.2.1 are:

• From Image2image in 2D to image2any: deepImageJ 2.1.10 runs models defined for 3D image processing, image classification or object detection through region proposal networks. It deals with miscellaneous output formats and shapes.
• GPU support.
• Compatibility with PyTorch: All PyTorch models stored as TorchScript can be for the first time loaded into ImageJ/Fiji through deepImageJ.
• New formats for image pre and post-processing: ImageJ macros (.txt/.ijm) and Java code&classes (.jar/.class) are supported and can be combined.
• Compatibility with the Bioimage Model Zoo🦒: deepImageJ is a consumer of the trained models in the BioImage Model Zoo
• Addition of a Validation plugin
• Support of headless mode for ImageJ/Fiji
• Full support of PyImageJ
• Add of test feature in DIJ Run

Changes with respect to 2.1.10:

• Allow fractional offsets (multiples of 0.5) to support stardist
• Upgrade Pytorch from 1.7 to 1.9

IMPORTANT (only for ImageJ1):

In order to run the plugin in GPU mode, the dependency libtensorflow_jni-1.15.0 has to be substituted by libtensorflow_jni_gpu-1.15.0. There should only be one of those two dependencies installed in the ImageJ1 directory.
DeepImageJ for GPU processing works even if there are is not any GPU available. However, the GPU version DOES NOT WORK on macOS distributions.

Content of DeepImageJ_dependencies_2110.zip:

• libtensorflow-1.15.0.jar
• libtensorflow_jni-1.15.0.jar
• proto-1.15.0.jar
• protobuf-java-3.5.1.jar
• junit-4.12.jar
• hamcrest-core-1.3.jar
• snakeyaml-1.21.jar
• slf4j-simple-1.7.26.jar
• slf4j-api-1.7.25.jar
• pytorch-native-auto-1.9.1.jar
• pytorch-engine-0.15.0.jar
• npy-0.3.3.jar
• kotlin-stdlib-1.3.72.jar
• jna-5.3.0.jar
• imagej-tensorflow-1.1.6.jar
• gson-2.8.5.jar
• commons-compress-1.20.jar
• api-0.15.0.jar

DeepImageJ-2.1.10

Release of DeepImageJ version 2.1.10.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

The main changes with respect to DeepImageJ 1.2.1 are:

• From Image2image in 2D to image2any: deepImageJ 2.1.10 runs models defined for 3D image processing, image classification or object detection through region proposal networks. It deals with miscellaneous output formats and shapes.
• GPU support.
• Compatibility with PyTorch: All PyTorch models stored as TorchScript can be for the first time loaded into ImageJ/Fiji through deepImageJ.
• New formats for image pre and post-processing: ImageJ macros (.txt/.ijm) and Java code&classes (.jar/.class) are supported and can be combined.
• Compatibility with the Bioimage Model Zoo🦒: deepImageJ is a consumer of the trained models in the BioImage Model Zoo
• Addition of a Validation plugin
• Support of headless mode for ImageJ/Fiji
• Full support of PyImageJ
• Add of test feature in DIJ Run

Changes with respect to 2.1.9:

• Updated the plugin to Bioimage.io 0.4.0
• Deprecate the model.yaml for an rdf.yaml
• Add DeepImageJ Test Button inside DeepImageJ Run
• Add support for PyImageJ

IMPORTANT (only for ImageJ1):

In order to run the plugin in GPU mode, the dependency libtensorflow_jni-1.15.0 has to be substituted by libtensorflow_jni_gpu-1.15.0. There should only be one of those two dependencies installed in the ImageJ1 directory.
DeepImageJ for GPU processing works even if there are is not any GPU available. However, the GPU version DOES NOT WORK on macOS distributions.

Content of DeepImageJ_dependencies_2110.zip:

• libtensorflow-1.15.0.jar
• libtensorflow_jni-1.15.0.jar
• proto-1.15.0.jar
• protobuf-java-3.5.1.jar
• junit-4.12.jar
• hamcrest-core-1.3.jar
• snakeyaml-1.21.jar
• slf4j-simple-1.7.26.jar
• slf4j-api-1.7.25.jar
• pytorch-native-auto-1.7.0.jar
• pytorch-engine-0.9.0.jar
• npy-0.3.3.jar
• kotlin-stdlib-1.3.72.jar
• jna-5.3.0.jar
• imagej-tensorflow-1.1.6.jar
• gson-2.8.5.jar
• commons-compress-1.20.jar
• api-0.9.0.jar

DeepImageJ-2.1.9

Release of DeepImageJ version 2.1.9.

To download already pretrained models for this version and for more information please go to:
https://deepimagej.github.io/deepimagej/models.html or https://bioimage.io/#/

The main changes with respect to DeepImageJ 1.2.1 are:

• From Image2image in 2D to image2any: deepImageJ 2.1.9 runs models defined for 3D image processing, image classification or object detection through region proposal networks. It deals with miscellaneous output formats and shapes.
• GPU support.
• Compatibility with PyTorch: All PyTorch models stored as TorchScript can be for the first time loaded into ImageJ/Fiji through deepImageJ.
• New formats for image pre and post-processing: ImageJ macros (.txt/.ijm) and Java code&classes (.jar/.class) are supported and can be combined.
• Compatibility with the Bioimage Model Zoo🦒: deepImageJ is a consumer of the trained models in the BioImage Model Zoo
• Addition of a Validation plugin
• Support of headless mode for ImageJ/Fiji (full support of pyimage is not ready yet)

Changes with respect to 2.1.8:

• Change the behaviour of the offset factor for the Channels axis. Previously, the offset would not affect the size of the final recostructed image (only the scale factor affected it) but now it does for the Channels axis.

IMPORTANT (only for ImageJ1):

In order to run the plugin in GPU mode, the dependency libtensorflow_jni-1.15.0 has to be substituted by libtensorflow_jni_gpu-1.15.0. There should only be one of those two dependencies installed in the ImageJ1 directory.
DeepImageJ for GPU processing works even if there are is not any GPU available. However, the GPU version DOES NOT WORK on macOS distributions.

Content of DeepImageJ_dependencies_217.zip:

• libtensorflow-1.15.0.jar
• libtensorflow_jni-1.15.0.jar
• proto-1.15.0.jar
• protobuf-java-3.5.1.jar
• junit-4.12.jar
• hamcrest-core-1.3.jar
• snakeyaml-1.21.jar
• slf4j-simple-1.7.26.jar
• slf4j-api-1.7.25.jar
• pytorch-native-auto-1.7.0.jar
• pytorch-engine-0.9.0.jar
• npy-0.3.3.jar
• kotlin-stdlib-1.3.72.jar
• jna-5.3.0.jar
• imagej-tensorflow-1.1.6.jar
• gson-2.8.5.jar
• commons-compress-1.20.jar
• api-0.9.0.jar