Datasets
Standard Dataset
Fall detection dataset for SFDM
- Citation Author(s):
- Submitted by:
- Jinmo Yang
- Last updated:
- Mon, 09/02/2024 - 20:11
- DOI:
- 10.21227/4hmh-sq38
- License:
105 Views- Categories:
- Keywords:
Abstract
This is the video dataset for SFDM paper. Only the first 30 seconds are to be used. The last 10 seconds are extended so that the trimming of the video does not make each clip end abruptly before 30 seconds.
File naming style: {case study number}{sequence}{clip number}{opt: falsely detected as other sequence}.webm
There are four cases: Normal 1, Fall 1, Normal 2, and Fall 2. Normal 1 is a trivial case of not fall. Fall 1 shows transitioning from normal state to fall state to indicate the person has fallen. Normal 2 shows recovery from fall, hence not fall. Fall 2 represents the person’s attempt at recovery but failing in the end, hence fall.
Normal 1: normal state for 30 seconds. Fall 1: normal state for 15 seconds and fall state for 15 seconds. Normal 2: normal state for 10 seconds, fall state for 10 seconds, and normal state for 10 seconds. Fall 2: normal state for 7 seconds, fall state for 8 seconds, normal state for 7 seconds, and fall state for 8 seconds.
individual file naming style: {case study number}{sequence}{clip number}{opt: falsely detected as other sequence}.webm
There are four cases: Normal 1, Fall 1, Normal 2, and Fall 2. Normal 1 is a trivial case of not fall. Fall 1 shows transitioning from normal state to fall state to indicate the person has fallen. Normal 2 shows recovery from fall, hence not fall. Fall 2 represents the person’s attempt at recovery but failing in the end, hence fall.
Normal 1: normal state for 30 seconds. Fall 1: normal state for 15 seconds and fall state for 15 seconds. Normal 2: normal state for 10 seconds, fall state for 10 seconds, and normal state for 10 seconds. Fall 2: normal state for 7 seconds, fall state for 8 seconds, normal state for 7 seconds, and fall state for 8 seconds.
Comments
Test