DSpace Repository

Visual Fall Detection for the Elderly Using Deep Learning on an Embedded Device

Show simple item record

dc.contributor.advisor Dailey, Matthew N.
dc.contributor.author Gurung, Anubinda
dc.contributor.other Ekpanyapong, Mongkol
dc.contributor.other Taparugssanagorn, Attaphongse
dc.date.accessioned 2020-08-14T06:20:25Z
dc.date.available 2020-08-14T06:20:25Z
dc.date.issued 2020-08
dc.identifier.citation APA en_US
dc.identifier.uri http://www.cs.ait.ac.th/xmlui/handle/123456789/984
dc.description.abstract The elderly population is increasing year by year and it won’t be long until they represent a majority of the population in many countries. This trend has been further bolstered by decreasing birth rates across the world. In this current scenario, taking care of the elderly people is difficult since care takers are not abundant. It is not feasible to have someone constantly monitor the elderly people through out their daily life. It is crucial to find a solution to ease the burden on care takers and automate some of their responsibilities. One risk factor for all elderly is a fall. It can be fatal to them. Even after recovery, it could cause chronic pain and take away their independence. Many suffer from mental stress after a fall. It affects them both physically and mentally. This thesis paper presents a Long Short Term Neural Network(LSTM) that uses visual cues, openpose keypoints, from camera feed to detect a fall. This model is deployed on a Jetson TX2 since it is cheap and can be placed in any elder person’s house without much difficulty. In pursuit of training the model to detect a fall, I run many experiments and describe the process of development. I attempt to improve the accuracy of the model when generalizing to an unknown camera view while also showing high accuracy after training on all camera views. I manage to run the LSTM and enable predictions on video and IP camera feed. en_US
dc.language.iso en en_US
dc.subject Deep Learning en_US
dc.title Visual Fall Detection for the Elderly Using Deep Learning on an Embedded Device en_US
dc.type Thesis en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record

Search DSpace


Advanced Search

Browse

My Account