Head Pose Estimation on Top of Haar-Like Face Detection: A Study Using the Kinect Sensor
Head pose estimation is a crucial initial task for human face analysis, which is employed in several computer vision systems, such as: facial expression recognition, head gesture recognition, yawn detection, etc. In this work, we propose a frame-based approach to estimate the head pose on top of the...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2015-08-01
|
| Series: | Sensors |
| Subjects: | |
| Online Access: | http://www.mdpi.com/1424-8220/15/9/20945 |
| _version_ | 1798003600451960832 |
|---|---|
| author | Anwar Saeed Ayoub Al-Hamadi Ahmed Ghoneim |
| author_facet | Anwar Saeed Ayoub Al-Hamadi Ahmed Ghoneim |
| author_sort | Anwar Saeed |
| collection | DOAJ |
| description | Head pose estimation is a crucial initial task for human face analysis, which is employed in several computer vision systems, such as: facial expression recognition, head gesture recognition, yawn detection, etc. In this work, we propose a frame-based approach to estimate the head pose on top of the Viola and Jones (VJ) Haar-like face detector. Several appearance and depth-based feature types are employed for the pose estimation, where comparisons between them in terms of accuracy and speed are presented. It is clearly shown through this work that using the depth data, we improve the accuracy of the head pose estimation. Additionally, we can spot positive detections, faces in profile views detected by the frontal model, that are wrongly cropped due to background disturbances. We introduce a new depth-based feature descriptor that provides competitive estimation results with a lower computation time. Evaluation on a benchmark Kinect database shows that the histogram of oriented gradients and the developed depth-based features are more distinctive for the head pose estimation, where they compare favorably to the current state-of-the-art approaches. Using a concatenation of the aforementioned feature types, we achieved a head pose estimation with average errors not exceeding 5:1; 4:6; 4:2 for pitch, yaw and roll angles, respectively. |
| first_indexed | 2024-04-11T12:10:19Z |
| format | Article |
| id | doaj.art-11b5bbccb53a4551a37efeec0c2b91a6 |
| institution | Directory Open Access Journal |
| issn | 1424-8220 |
| language | English |
| last_indexed | 2024-04-11T12:10:19Z |
| publishDate | 2015-08-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Sensors |
| spelling | doaj.art-11b5bbccb53a4551a37efeec0c2b91a62022-12-22T04:24:37ZengMDPI AGSensors1424-82202015-08-01159209452096610.3390/s150920945s150920945Head Pose Estimation on Top of Haar-Like Face Detection: A Study Using the Kinect SensorAnwar Saeed0Ayoub Al-Hamadi1Ahmed Ghoneim2Institute for Information Technology and Communications (IIKT), Otto-von-Guericke-University Magdeburg, Magdeburg D-39016, GermanyInstitute for Information Technology and Communications (IIKT), Otto-von-Guericke-University Magdeburg, Magdeburg D-39016, GermanyDepartment of Software Engineering, College of Computer Science and Information Sciences, King Saud University, Riyadh 11451, Saudi ArabiaHead pose estimation is a crucial initial task for human face analysis, which is employed in several computer vision systems, such as: facial expression recognition, head gesture recognition, yawn detection, etc. In this work, we propose a frame-based approach to estimate the head pose on top of the Viola and Jones (VJ) Haar-like face detector. Several appearance and depth-based feature types are employed for the pose estimation, where comparisons between them in terms of accuracy and speed are presented. It is clearly shown through this work that using the depth data, we improve the accuracy of the head pose estimation. Additionally, we can spot positive detections, faces in profile views detected by the frontal model, that are wrongly cropped due to background disturbances. We introduce a new depth-based feature descriptor that provides competitive estimation results with a lower computation time. Evaluation on a benchmark Kinect database shows that the histogram of oriented gradients and the developed depth-based features are more distinctive for the head pose estimation, where they compare favorably to the current state-of-the-art approaches. Using a concatenation of the aforementioned feature types, we achieved a head pose estimation with average errors not exceeding 5:1; 4:6; 4:2 for pitch, yaw and roll angles, respectively.http://www.mdpi.com/1424-8220/15/9/20945head poselocal binary patternhistogram of gradientGabor filterKinect sensorsupport vector machineregression |
| spellingShingle | Anwar Saeed Ayoub Al-Hamadi Ahmed Ghoneim Head Pose Estimation on Top of Haar-Like Face Detection: A Study Using the Kinect Sensor Sensors head pose local binary pattern histogram of gradient Gabor filter Kinect sensor support vector machine regression |
| title | Head Pose Estimation on Top of Haar-Like Face Detection: A Study Using the Kinect Sensor |
| title_full | Head Pose Estimation on Top of Haar-Like Face Detection: A Study Using the Kinect Sensor |
| title_fullStr | Head Pose Estimation on Top of Haar-Like Face Detection: A Study Using the Kinect Sensor |
| title_full_unstemmed | Head Pose Estimation on Top of Haar-Like Face Detection: A Study Using the Kinect Sensor |
| title_short | Head Pose Estimation on Top of Haar-Like Face Detection: A Study Using the Kinect Sensor |
| title_sort | head pose estimation on top of haar like face detection a study using the kinect sensor |
| topic | head pose local binary pattern histogram of gradient Gabor filter Kinect sensor support vector machine regression |
| url | http://www.mdpi.com/1424-8220/15/9/20945 |
| work_keys_str_mv | AT anwarsaeed headposeestimationontopofhaarlikefacedetectionastudyusingthekinectsensor AT ayoubalhamadi headposeestimationontopofhaarlikefacedetectionastudyusingthekinectsensor AT ahmedghoneim headposeestimationontopofhaarlikefacedetectionastudyusingthekinectsensor |