Integrate senselab for speech analysis in health

[fabiocat93]

senselab is an open-source Python package for standardized voice and speech analysis, with a focus on health and clinical research. It provides:

• Preprocessing (VAD, resampling, normalization)
• Feature extraction (hand-crafted features such as eGeMAPS, and deep embeddings like WavLM/ECAPA)
• Utilities for reproducible pipelines in speech/voice biomarker studies

Integrating senselab into PyHealth would enable end-to-end workflows for voice and speech biomarkers, complementing existing EHR and sensor modules. This would extend PyHealth to support use cases such as depression detection, Parkinson’s monitoring, rehabilitation progress tracking, and cognitive health assessment through voice

[jhnwu3]

This is a cool idea.

Not sure, if it's easy for you to help port these things over @fabiocat93 into the processors/ directory. https://github.com/sunlabuiuc/PyHealth/tree/master/pyhealth/processors

We should be able to write some simple wrappers around processing samples with your audio specific transforms before converting to Torch.

I know @plandes has made a way to build dependency trees for specialized versions of pyhealth.