Fall Detection in Elevator Cages Based on XGBoost and LSTM

Caorong Xu, Yuanping Xu, Zhijie Xu, Benjun Guo, Chaolong Zhang, Jian Huang, Xin Deng

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

4 Citations (Scopus)

Abstract

Fall detection has always been challenging. There are many current works that have achieved good results in fall detections, but most of the existing works only consider single feature dimension (i.e., temporal features or spatial features). This study proposes a method that combines LSTM and XGBoost model to detect human body falls through fusing both temporal and spatial features. It starts with extracting bond points on human body in each frame of the video by using AlphaPose. In the second step, three features of the human body (i.e., vertical height, aspect ratio of the human external rectangle and angle of the knee joint) are calculated as spatial features through the extracted bone point information. Then use LSTM to learn these three kinds of features on both the temporal and spatial dimension. Furthermore, this model integrates the XGBoost to learn multiple features to improve the recognition rate. Finally, various human body fall detections in elevator cabs are applied to test the usability and validity of this study. The experimental results of the whole model reach 92.11% recognition accuracy, and 93.33% on the the F1-measure index.

Original languageEnglish
Title of host publication2021 26th International Conference on Automation and Computing
Subtitle of host publicationSystem Intelligence through Automation and Computing, ICAC 2021
EditorsChenguang Yang
PublisherIEEE
Number of pages6
ISBN (Electronic)9781860435577
ISBN (Print)9781665443524
DOIs
Publication statusPublished - 15 Nov 2021
Event26th International Conference on Automation and Computing - University of Portsmouth, Portsmouth, United Kingdom
Duration: 2 Sep 20214 Sep 2021
Conference number: 26
http://www.cacsuk.co.uk/index.php/icac2021
https://www.ieee-ras.org/conferences-workshops/technically-co-sponsored/icac
https://ieeexplore.ieee.org/xpl/conhome/9594055/proceeding

Conference

Conference26th International Conference on Automation and Computing
Abbreviated titleICAC 2021
Country/TerritoryUnited Kingdom
CityPortsmouth
Period2/09/214/09/21
Internet address

Fingerprint

Dive into the research topics of 'Fall Detection in Elevator Cages Based on XGBoost and LSTM'. Together they form a unique fingerprint.

Cite this