A Quantitative Comparison of Overlapping and Non-Overlapping Sliding Windows for Human Activity Recognition Using Inertial Sensors
The sliding window technique is widely used to segment inertial sensor signals, i.e., accelerometers and gyroscopes, for activity recognition. In this technique, the sensor signals are partitioned into fix sized time windows which can be of two types: (1) non-overlapping windows, in which time windo...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-11-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1424-8220/19/22/5026 |
| _version_ | 1811307207599325184 |
|---|---|
| author | Akbar Dehghani Omid Sarbishei Tristan Glatard Emad Shihab |
| author_facet | Akbar Dehghani Omid Sarbishei Tristan Glatard Emad Shihab |
| author_sort | Akbar Dehghani |
| collection | DOAJ |
| description | The sliding window technique is widely used to segment inertial sensor signals, i.e., accelerometers and gyroscopes, for activity recognition. In this technique, the sensor signals are partitioned into fix sized time windows which can be of two types: (1) non-overlapping windows, in which time windows do not intersect, and (2) overlapping windows, in which they do. There is a generalized idea about the positive impact of using overlapping sliding windows on the performance of recognition systems in Human Activity Recognition. In this paper, we analyze the impact of overlapping sliding windows on the performance of Human Activity Recognition systems with different evaluation techniques, namely, subject-dependent cross validation and subject-independent cross validation. Our results show that the performance improvements regarding overlapping windowing reported in the literature seem to be associated with the underlying limitations of subject-dependent cross validation. Furthermore, we do not observe any performance gain from the use of such technique in conjunction with subject-independent cross validation. We conclude that when using subject-independent cross validation, non-overlapping sliding windows reach the same performance as sliding windows. This result has significant implications on the resource usage for training the human activity recognition systems. |
| first_indexed | 2024-04-13T09:00:30Z |
| format | Article |
| id | doaj.art-e8ffcda0e08545bcb490b463faf194f6 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-13T09:00:30Z |
| publishDate | 2019-11-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-e8ffcda0e08545bcb490b463faf194f62022-12-22T02:53:10ZengMDPI AGSensors1424-82202019-11-011922502610.3390/s19225026s19225026A Quantitative Comparison of Overlapping and Non-Overlapping Sliding Windows for Human Activity Recognition Using Inertial SensorsAkbar Dehghani0Omid Sarbishei1Tristan Glatard2Emad Shihab3Department of Computer Science and Software Engineering, Concordia University, Montreal, QC H3G 1M8, CanadaResearch and Development Department, Motsai Research, Saint Bruno, QC J3V 6B7, CanadaDepartment of Computer Science and Software Engineering, Concordia University, Montreal, QC H3G 1M8, CanadaDepartment of Computer Science and Software Engineering, Concordia University, Montreal, QC H3G 1M8, CanadaThe sliding window technique is widely used to segment inertial sensor signals, i.e., accelerometers and gyroscopes, for activity recognition. In this technique, the sensor signals are partitioned into fix sized time windows which can be of two types: (1) non-overlapping windows, in which time windows do not intersect, and (2) overlapping windows, in which they do. There is a generalized idea about the positive impact of using overlapping sliding windows on the performance of recognition systems in Human Activity Recognition. In this paper, we analyze the impact of overlapping sliding windows on the performance of Human Activity Recognition systems with different evaluation techniques, namely, subject-dependent cross validation and subject-independent cross validation. Our results show that the performance improvements regarding overlapping windowing reported in the literature seem to be associated with the underlying limitations of subject-dependent cross validation. Furthermore, we do not observe any performance gain from the use of such technique in conjunction with subject-independent cross validation. We conclude that when using subject-independent cross validation, non-overlapping sliding windows reach the same performance as sliding windows. This result has significant implications on the resource usage for training the human activity recognition systems.https://www.mdpi.com/1424-8220/19/22/5026activity recognitioninertial sensorssupervised classification |
| spellingShingle | Akbar Dehghani Omid Sarbishei Tristan Glatard Emad Shihab A Quantitative Comparison of Overlapping and Non-Overlapping Sliding Windows for Human Activity Recognition Using Inertial Sensors Sensors activity recognition inertial sensors supervised classification |
| title | A Quantitative Comparison of Overlapping and Non-Overlapping Sliding Windows for Human Activity Recognition Using Inertial Sensors |
| title_full | A Quantitative Comparison of Overlapping and Non-Overlapping Sliding Windows for Human Activity Recognition Using Inertial Sensors |
| title_fullStr | A Quantitative Comparison of Overlapping and Non-Overlapping Sliding Windows for Human Activity Recognition Using Inertial Sensors |
| title_full_unstemmed | A Quantitative Comparison of Overlapping and Non-Overlapping Sliding Windows for Human Activity Recognition Using Inertial Sensors |
| title_short | A Quantitative Comparison of Overlapping and Non-Overlapping Sliding Windows for Human Activity Recognition Using Inertial Sensors |
| title_sort | quantitative comparison of overlapping and non overlapping sliding windows for human activity recognition using inertial sensors |
| topic | activity recognition inertial sensors supervised classification |
| url | https://www.mdpi.com/1424-8220/19/22/5026 |
| work_keys_str_mv | AT akbardehghani aquantitativecomparisonofoverlappingandnonoverlappingslidingwindowsforhumanactivityrecognitionusinginertialsensors AT omidsarbishei aquantitativecomparisonofoverlappingandnonoverlappingslidingwindowsforhumanactivityrecognitionusinginertialsensors AT tristanglatard aquantitativecomparisonofoverlappingandnonoverlappingslidingwindowsforhumanactivityrecognitionusinginertialsensors AT emadshihab aquantitativecomparisonofoverlappingandnonoverlappingslidingwindowsforhumanactivityrecognitionusinginertialsensors AT akbardehghani quantitativecomparisonofoverlappingandnonoverlappingslidingwindowsforhumanactivityrecognitionusinginertialsensors AT omidsarbishei quantitativecomparisonofoverlappingandnonoverlappingslidingwindowsforhumanactivityrecognitionusinginertialsensors AT tristanglatard quantitativecomparisonofoverlappingandnonoverlappingslidingwindowsforhumanactivityrecognitionusinginertialsensors AT emadshihab quantitativecomparisonofoverlappingandnonoverlappingslidingwindowsforhumanactivityrecognitionusinginertialsensors |