u net fusion rfi

We have hosted the application u net fusion rfi in order to run this application in our online workstations with Wine or directly.

Quick description about u net fusion rfi:

See original code here: https: //github.com/jakeret/tf_unet

Currently this project is based on Tensorflow 1.13 code base and there are no plans to transfer to TF version 2. The primary improvements to this code base include a training and evaluation framework, along with a fusion based approach to detection, combining a number of models (currently hard coded to two trained models) along with Sum Threshold as an additional "expert." Additional work is being done to add custom layers to this model for further experimentation, including Squeeze/Excitation layers (unimplemented.)

Sum Threshold (in fusion as an expert, and in testing as a comparison) requires the use of AOFlagger by Andre Offringa. You can find this code at https: //gitlab.com/aroffringa/aoflagger. This project will use the aoflagger program within the code, so you may need to ensure that any environment variables are set for aoflagger before use.

cite: https: //sourceforge.net/p/u-net-fusion-rfi/wiki/cite/.

Features:

• Real spectrogram images, labeled by Sum Threshold included in files
• Synthetically generated test spectograms included
• Ability to fuse model results with Sum Threshold for prediction evaluation

Audience: Education.
User interface: Command-line.
Programming Language: Python.
Categories: