From 7cb16a1276c6a250017ce7d31afb29b5b40d0e39 Mon Sep 17 00:00:00 2001 From: Shing Chan Date: Fri, 20 Dec 2024 10:47:44 -0300 Subject: [PATCH] improve README.md --- README.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 760f8b5..6252d64 100644 --- a/README.md +++ b/README.md @@ -1,8 +1,8 @@ # stepcount -A step-counting model based on self-supervised learning for wrist-worn accelerometer data. +Improved step counting based on a foundation model for wrist-worn accelerometers. -The SSL model was pre-trained using the large-scale [UK Biobank Accelerometer Dataset](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0169649), and fine-tuned on the [OxWalk Dataset](https://ora.ox.ac.uk/objects/uuid:19d3cb34-e2b3-4177-91b6-1bad0e0163e7). +The foundation model was trained using self-supervised learning on the large-scale [UK Biobank Accelerometer Dataset](https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0169649), and fine-tuned on the [OxWalk Dataset](https://ora.ox.ac.uk/objects/uuid:19d3cb34-e2b3-4177-91b6-1bad0e0163e7). The command-line tool can process Axivity AX3 files (UK Biobank, China Kadoorie Biobank) directly. For consumer devices like Fitbit and Apple Watch, convert them to raw CSV first.