Otsu Image Threshold Broken in tfjs-react-native

[mattsoulanille]

Please make sure that this is a bug. As per our
GitHub Policy,
we only address code/doc bugs, performance issues, feature requests and
build/installation issues on GitHub. tag:bug_template

System information

• Have I written custom code (as opposed to using a stock example script provided in TensorFlow.js): No
• OS Platform and Distribution (e.g., Linux Ubuntu 16.04): N/A
• Mobile device (e.g. iPhone 8, Pixel 2, Samsung Galaxy) if the issue happens on mobile device: Android 9.0 Samsung Galaxy S9 Plus
• TensorFlow.js installed from (npm or script link): master branch
• TensorFlow.js version (use command below): master branch
• Browser version: N/A
• Tensorflow.js Converter Version: N/A

Describe the current behavior
In the image threshold op, otsu's method returns a different (incorrect) result for webgl in tfjs-react-native than in the browser.
Describe the expected behavior
Otsu's method returns the correct result in tfjs-react-native.

Standalone code to reproduce the issue
Provide a reproducible test case that is the bare minimum necessary to generate
the problem. If possible, please share a link to Colab/CodePen/any notebook.
Run CI tests on tfjs-react-native. The bug may also appear when running the integration tests of tfjs-react-native on a local android phone.

Other info / logs Include any logs or source code that would be helpful to
diagnose the problem. If including tracebacks, please include the full
traceback. Large logs and files should be attached.

I visualized the test results and for some reason, the otsu threshold test output matches the expected output for binary. This is almost certainly a coincidence.

Logs:
https://console.cloud.google.com/cloud-build/builds/15aa4850-a102-45f4-9525-c06588d6f668;step=7?project=learnjs-174218

More details on this PR: #5217

[mattsoulanille]

Reference colab contributed to the TF repo

TF Proposal