sdimi/Step2heart
🏋️ Extract generalizable physiological representations from sensor time-series data with self-supervision (code for CHIL 2021, NeurIPS-W 2020 papers)
This project helps health researchers and data scientists predict high-level health outcomes like cardiorespiratory fitness (VO2max) or demographic characteristics from everyday wearable sensor data. It takes in raw activity data (like acceleration from a smartwatch) and uses it to estimate heart rate, which then helps create meaningful physiological representations. These representations can be used to understand an individual's health and lifestyle.
No commits in the last 6 months.
Use this if you have large datasets of wearable activity sensor data and are looking for a way to extract generalizable physiological insights to predict health or fitness markers, especially when detailed heart rate data or labeled health outcomes are scarce.
Not ideal if you need to analyze raw ECG signals directly, are working with limited or proprietary sensor types, or primarily interested in real-time heart rate monitoring rather than long-term health predictions.
Stars
19
Forks
2
Language
Python
License
GPL-3.0
Category
Last pushed
Sep 07, 2023
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/embeddings/sdimi/Step2heart"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
mims-harvard/ClinVec
ClinVec: Unified Embeddings of Clinical Codes Enable Knowledge-Grounded AI in Medicine
NYUMedML/DeepEHR
Chronic Disease Prediction Using Medical Notes
mims-harvard/SHEPHERD
SHEPHERD: Few shot learning for phenotype-driven diagnosis of patients with rare genetic diseases
biocentral/biocentral_server
Compute functionality for biocentral.
nomic-ai/contrastors
Train Models Contrastively in Pytorch