Action Recognition Using Single-Pixel Time-of-Flight Detection
Action recognition is a challenging task that plays an important role in many robotic systems, which highly depend on visual input feeds. However, due to privacy concerns, it is important to find a method which can recognise actions without using visual feed. In this paper, we propose a concept for...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2019-04-01
|
| Series: | Entropy |
| Subjects: | |
| Online Access: | https://www.mdpi.com/1099-4300/21/4/414 |
| _version_ | 1811262829865467904 |
|---|---|
| author | Ikechukwu Ofodile Ahmed Helmi Albert Clapés Egils Avots Kerttu Maria Peensoo Sandhra-Mirella Valdma Andreas Valdmann Heli Valtna-Lukner Sergey Omelkov Sergio Escalera Cagri Ozcinar Gholamreza Anbarjafari |
| author_facet | Ikechukwu Ofodile Ahmed Helmi Albert Clapés Egils Avots Kerttu Maria Peensoo Sandhra-Mirella Valdma Andreas Valdmann Heli Valtna-Lukner Sergey Omelkov Sergio Escalera Cagri Ozcinar Gholamreza Anbarjafari |
| author_sort | Ikechukwu Ofodile |
| collection | DOAJ |
| description | Action recognition is a challenging task that plays an important role in many robotic systems, which highly depend on visual input feeds. However, due to privacy concerns, it is important to find a method which can recognise actions without using visual feed. In this paper, we propose a concept for detecting actions while preserving the test subject’s privacy. Our proposed method relies only on recording the temporal evolution of light pulses scattered back from the scene. Such data trace to record one action contains a sequence of one-dimensional arrays of voltage values acquired by a single-pixel detector at 1 GHz repetition rate. Information about both the distance to the object and its shape are embedded in the traces. We apply machine learning in the form of recurrent neural networks for data analysis and demonstrate successful action recognition. The experimental results show that our proposed method could achieve on average <inline-formula> <math display="inline"> <semantics> <mrow> <mn>96.47</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula> accuracy on the actions walking forward, walking backwards, sitting down, standing up and waving hand, using recurrent neural network. |
| first_indexed | 2024-04-12T19:33:49Z |
| format | Article |
| id | doaj.art-21039b779fba4e4e9fbbd6c7353a4987 |
| institution | Directory Open Access Journal |
| issn | 1099-4300 |
| language | English |
| last_indexed | 2024-04-12T19:33:49Z |
| publishDate | 2019-04-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Entropy |
| spelling | doaj.art-21039b779fba4e4e9fbbd6c7353a49872022-12-22T03:19:16ZengMDPI AGEntropy1099-43002019-04-0121441410.3390/e21040414e21040414Action Recognition Using Single-Pixel Time-of-Flight DetectionIkechukwu Ofodile0Ahmed Helmi1Albert Clapés2Egils Avots3Kerttu Maria Peensoo4Sandhra-Mirella Valdma5Andreas Valdmann6Heli Valtna-Lukner7Sergey Omelkov8Sergio Escalera9Cagri Ozcinar10Gholamreza Anbarjafari11iCv Lab, Institute of Technology, University of Tartu, 50411 Tartu, EstoniaiCv Lab, Institute of Technology, University of Tartu, 50411 Tartu, EstoniaUniversity of Barcelona, 08007 Barcelona, SpainiCv Lab, Institute of Technology, University of Tartu, 50411 Tartu, EstoniaInstitute of Physics, University of Tartu, 50411 Tartu, EstoniaInstitute of Physics, University of Tartu, 50411 Tartu, EstoniaInstitute of Physics, University of Tartu, 50411 Tartu, EstoniaInstitute of Physics, University of Tartu, 50411 Tartu, EstoniaInstitute of Physics, University of Tartu, 50411 Tartu, EstoniaUniversity of Barcelona, 08007 Barcelona, SpainTrinity College Dublin, Dublin 2, IrelandiCv Lab, Institute of Technology, University of Tartu, 50411 Tartu, EstoniaAction recognition is a challenging task that plays an important role in many robotic systems, which highly depend on visual input feeds. However, due to privacy concerns, it is important to find a method which can recognise actions without using visual feed. In this paper, we propose a concept for detecting actions while preserving the test subject’s privacy. Our proposed method relies only on recording the temporal evolution of light pulses scattered back from the scene. Such data trace to record one action contains a sequence of one-dimensional arrays of voltage values acquired by a single-pixel detector at 1 GHz repetition rate. Information about both the distance to the object and its shape are embedded in the traces. We apply machine learning in the form of recurrent neural networks for data analysis and demonstrate successful action recognition. The experimental results show that our proposed method could achieve on average <inline-formula> <math display="inline"> <semantics> <mrow> <mn>96.47</mn> <mo>%</mo> </mrow> </semantics> </math> </inline-formula> accuracy on the actions walking forward, walking backwards, sitting down, standing up and waving hand, using recurrent neural network.https://www.mdpi.com/1099-4300/21/4/414single pixel single photon image acquisitiontime-of-flightaction recognition |
| spellingShingle | Ikechukwu Ofodile Ahmed Helmi Albert Clapés Egils Avots Kerttu Maria Peensoo Sandhra-Mirella Valdma Andreas Valdmann Heli Valtna-Lukner Sergey Omelkov Sergio Escalera Cagri Ozcinar Gholamreza Anbarjafari Action Recognition Using Single-Pixel Time-of-Flight Detection Entropy single pixel single photon image acquisition time-of-flight action recognition |
| title | Action Recognition Using Single-Pixel Time-of-Flight Detection |
| title_full | Action Recognition Using Single-Pixel Time-of-Flight Detection |
| title_fullStr | Action Recognition Using Single-Pixel Time-of-Flight Detection |
| title_full_unstemmed | Action Recognition Using Single-Pixel Time-of-Flight Detection |
| title_short | Action Recognition Using Single-Pixel Time-of-Flight Detection |
| title_sort | action recognition using single pixel time of flight detection |
| topic | single pixel single photon image acquisition time-of-flight action recognition |
| url | https://www.mdpi.com/1099-4300/21/4/414 |
| work_keys_str_mv | AT ikechukwuofodile actionrecognitionusingsinglepixeltimeofflightdetection AT ahmedhelmi actionrecognitionusingsinglepixeltimeofflightdetection AT albertclapes actionrecognitionusingsinglepixeltimeofflightdetection AT egilsavots actionrecognitionusingsinglepixeltimeofflightdetection AT kerttumariapeensoo actionrecognitionusingsinglepixeltimeofflightdetection AT sandhramirellavaldma actionrecognitionusingsinglepixeltimeofflightdetection AT andreasvaldmann actionrecognitionusingsinglepixeltimeofflightdetection AT helivaltnalukner actionrecognitionusingsinglepixeltimeofflightdetection AT sergeyomelkov actionrecognitionusingsinglepixeltimeofflightdetection AT sergioescalera actionrecognitionusingsinglepixeltimeofflightdetection AT cagriozcinar actionrecognitionusingsinglepixeltimeofflightdetection AT gholamrezaanbarjafari actionrecognitionusingsinglepixeltimeofflightdetection |