Asset Details
Do you wish to reserve the book?
Unusual Activity Recognition Based on 2D-Skeleton Data
Hey, we have placed the reservation for you!
By the way, why not check out events that you can attend while you pick your title.
Oops! Something went wrong.
Looks like we were not able to place the reservation. Kindly try again later.
Are you sure you want to remove the book from the shelf?
Unusual Activity Recognition Based on 2D-Skeleton Data
Do you wish to request the book?
Unusual Activity Recognition Based on 2D-Skeleton Data
Please be aware that the book you have requested cannot be checked out. If you would like to checkout this book, you can reserve another copy
We have requested the book for you!
Your request is successful and it will be processed during the Library working hours. Please check the status of your request in My Requests.
Oops! Something went wrong.
Looks like we were not able to place your request. Kindly try again later.
Journal Article
Unusual Activity Recognition Based on 2D-Skeleton Data
2026
Overview
In some hospitals or similar institutions, it can be helpful to anonymously monitor patients’ activities so that self-harm and violence against others can be detected and prevented early. The ISAS 2025 Challenge addresses this problem by looking for the best method to classify eight predefined types of human activities based on 2D-skeleton data. To solve this problem, we consider two different approaches: In our first approach, we apply TinyHAR directly on the given 2D-skeleton data and test different feature scaling techniques and a body-centered coordinate system with different origins. In our second approach, we transform the 2D-skeleton data back into a video sequence and use these videos as input for video classification models. Finally, we state the Leave-One-Subject-Out (LOSO) mean and standard deviation of accuracy and macro F 1 -Score. Since our chosen models vary a lot in the number of parameters and size, we also state the power consumption in Wh and the VRAM usage of the GPU.
Publisher
IOP Publishing
Subject
