List view
We want to make sure that we have as many avenues as possible to test out automated segmentation due to the large breadth of environments we will encounter in the passive acoustic monitoring field. The idea here is that we will add in an extra parameter into the isolation parameters called "model" that allows a user to select which neural network frameworks will be deployed to attempt to extract bird vocalizations from audio data.
No due date•7/7 issues closedCreating pipeline to gauge how long each function is taking to identify bottlenecks and then using data structures and linear algebra to improve temporal performance.
No due date•4/4 issues closedMake it so that users can tweak different variables on the Isolate function so that they can try and get closer to the human labels. This involves deploying different techniques such as smoothing out the local scores arrays as they are passed in, modifying the thresholds in different ways, as well as developing entirely new call isolation techniques.
No due date•6/7 issues closed- No due date•9/13 issues closed
This involves converting the jupyter notebook into readme.md as well as talking about the general purpose of the package. Include docstrings that describe functions with respect to their purpose in relation to other functions, as well as the inputs and outputs.
No due date•3/4 issues closed