Skip to main content


Digital Biomaker Library.

Advanced Digital Biomaker

OpenDBM is a software package that allows for calculation of digital biomarkers from video/audio of a person's behavior by combining tools to measure behavioral characteristics like facial activity, voice, and movement into a single package for measurement of overall behavior.

Ease of use. OpenDBM is designed for ease of use, expanding the availability of such tools to the wider scientific community.

Written in Python—support all OS platforms

It using Python language so you can integrate with other biomaker or ML libraries

Many OS platforms, one code. You can focus on building researching medical purpose and the single codebase can share code across platforms. With OpenDBM, one team can maintain multiple OS platforms and share a common researching application.

from opendbm.facial import FacialActivity

model = FacialActivity()
landmark = model.get_landmark()

Rich Features for Researchers

Through OpenDBM, a user can objectively and sensitively measure behavioral characteristics such as facial activity, vocal acoustics, characteristics of speech, patterns of movement, and oculomotion.

From those behavioral characteristics, they can measure clinically meaningful symptomatology such as emotional expressivity, prosody of voice, valence of speech, and severity of tremor––among many others.

Talks and Videos

We’ve recorded some instructional videos (listed and linked to within the documentation) that should help the user get through common steps such as installation, usage, etc.

The OpenDBM Contributor Team has put together a short video where they explained about the installation in some platforms.

Community Driven

AiCure released OpenDBM in 2020 and has been maintaining it ever since.

We want to ensure that OpenDBM feels like a community of like-minded researchers and clinicians. Hence, there are a few ways we encourage users to stay involved––and why we encourage you to join DiME, too! Most importantly, if you’re interested in OpenDBM, star the repo and sign up for our listserv for all updates.

If you’re thinking about contributing to OpenDBM––to which we say kudos––please reach out to us. We’ve written code of conduct and contribution guidelines but also want to do whatever we can to help.

A point that was mentioned earlier and cannot be emphasized enough is that OpenDBM is simply a compilation of existing but disparate open-source software tools that we’ve built on top of. All these tools are of course listed in the OpenDBM dependencies but we want to acknowledge just a few here: OpenFace, built on OpenCV, is at the heart of all facial measures and even some of the movement ones. Parselmouth and its reliance on the Praat software library lies behind most of the vocal acoustic measures. DeepSpeech was used for all speech transcription and NLTK is utilized for a lot of language metrics. OpenDBM would not be possible without these––and several other––open source software packages.

  • OpenFace
  • Parselmouth
  • DeepSpeech

Give it a try

  1. Run this

    pip install opendbm
  2. Read these

    Get startedLearn basics