From f01646ce46375ff64ea0a64423a3788f787af3c7 Mon Sep 17 00:00:00 2001 From: Timothy Sipkens Date: Thu, 19 Sep 2024 12:23:43 -0400 Subject: [PATCH] Update README.md Add updated links to ONNX file. --- README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 4b7d72c..5449791 100644 --- a/README.md +++ b/README.md @@ -23,7 +23,7 @@ The program is primarily composed of two analysis packages, which will be discus 2. **+pp** - which determines primarily particle information, often from the binary image generated by the methods in the **agg** package noted above. -This code also includes a set of utility functions in the **+tools** package and Python code necessary to implement a convolutional neural network used for segmentation in the [carboseg](https://github.com/tsipkens/atems/tree/master/carboseg) folder. +This code also includes a set of utility functions in the **+tools** package and Python code necessary to implement a convolutional neural network used for segmentation in the [carboseg](https://github.com/tsipkens/atems/tree/master/carboseg) folder. The corresponds ONNX file is temporarily available here: https://github.com/tsipkens/FPN-resnet50-imagenet. ### In this README @@ -227,7 +227,7 @@ Though, the technique still occasionally fails, particularly if the function doe ### + **carboseg** / Neural network-based segmentation -This `seg_carboseg(...)` function employs Python to implement a convolutional neural network (CNN) for segmentation as described by [Sipkens et al.][ptech.cnn] Details and code for the training of the network are available in a parallel repository at https://github.com/maxfrei750/CarbonBlackSegmentation, with primary contributions by Max Frei ([@maxfrei750](https://github.com/maxfrei750)). The implementation here makes use of the ONNX file output (to be downloaded [here](https://uni-duisburg-essen.sciebo.de/s/J7bS47nZadg4bBH/download)) from that procedure and employs the Python ONNX runtime for execution. Use of this function requires the necessary Python environment as a pre-requisite. +This `seg_carboseg(...)` function employs Python to implement a convolutional neural network (CNN) for segmentation as described by [Sipkens et al.][ptech.cnn] Details and code for the training of the network are available in a parallel repository at https://github.com/maxfrei750/CarbonBlackSegmentation, with primary contributions by Max Frei ([@maxfrei750](https://github.com/maxfrei750)). The implementation here makes use of the ONNX file output (to be downloaded [here](https://github.com/tsipkens/FPN-resnet50-imagenet)) from that procedure and employs the Python ONNX runtime for execution. Use of this function requires the necessary Python environment as a pre-requisite. > We also note that, as of this writing, MATLAB does not support the necessary layers to import the ONNX as a native MATLAB object.