Human behavior understanding through 3D data

Download

index.pdf

Date

2015

Author

Akdağ, Erkut

Metadata

Show full item record

Item Usage Stats

127
views

50
downloads

In the human action recognition area, so far 2D action recognition has been studied extensively. Recently some studies, understanding human actions in 3D is emerging due to development of devices collecting 3D data. In this thesis, a new human behavior recognition method, that we call silhouette flows, is proposed for 3D data sequences of depth map. The method proposed in this thesis constitutes two steps, which are the feature extraction and classification. In feature extraction part, motion features are extracted from the 3D binary depth data in order to discern possibilities for action within the environment. For this purpose, the 3D depth data is projected on to cartesian planes in order to obtain silhouettes in frontal, top and side views and then optical flow vector fields on these planes over each frame of the video are computed. After finding these flow vectors, averages are prepared according to the motion vector values separately for negative and positive values for each frame of each plane. In order to recognize various human behaviors, each frame in video is divided into some meaningful blocks. According to the significant motion blocks, the final motion feature vector is obtained. Then, this motion feature vector is given to the SVM classification system and the results are investigated. All experiments are conducted on depth map data “MSR Action3D Dataset”. This dataset includes twenty human actions depth map sequences recorded with Microsoft Kinect depth sensors for ten different people. The experimental results are quite successful and the proposed method outperformed in some test the other methods existing in literature for the same data.

Subject Keywords

Three-dimensional imaging., Imaging systems., Human mechanics., Image processing., Optical data processing.

URI

http://etd.lib.metu.edu.tr/upload/12619581/index.pdf
https://hdl.handle.net/11511/25309

Collections

Graduate School of Natural and Applied Sciences, Thesis

Suggestions

OpenMETU
Core

Human action recognition for various input characteristics using 3 dimensional residual networks Tüfekci, Gülin; Ulusoy, İlkay; Department of Electrical and Electronics Engineering (2019) Action recognition using deep neural networks is a far-reaching research area which has been commonly utilized in applications such as statistical analysis of human behavior, detecting abnormalities using surveillance cameras and robotic systems. Previous studies have been performing researches to propose new machine learning algorithms and deep network architectures to obtain higher recognition accuracy levels. Instead of suggesting a network resulting in small accuracy gain, this thesis focuses on evaluat...
Object recognition and segmentation via shape models Altınoklu, Metin Burak; Ulusoy, İlkay; Tarı, Zehra Sibel; Department of Electrical and Electronics Engineering (2016) In this thesis, the problem of object detection, recognition and segmentation in computer vision is addressed with shape based methods. An efficient object detection method based on a sparse skeleton has been proposed. The proposed method is an improved chamfer template matching method for recognition of articulated objects. Using a probabilistic graphical model structure, shape variation is represented in a skeletal shape model, where nodes correspond to parts consisting of lines and edges correspond to pa...
Superpixel based image sequence representation and motion estimation İnce, Kutalmış Gökalp; Alatan, Abdullah Aydın; Demirekler, Mübeccel; Department of Electrical and Electronics Engineering (2017) In this study a superpixel based representation of image sequences is proposed. For superpixel extraction, a novel gradient ascent approach, in which spatial and spectral statistics are utilized to obtain an optimal Bayesian classiﬁer for pixel to superpixel label assignment, is proposed. Utilization of the spectral and spatial statistics reduce the dependency on user selected global parameters, while increasing the robustness and adaptability. Proposed Local Adaptive Superpixels (LASP) approach exploits he...
Hierarchical representations for visual object tracking by detection Beşbınar, Beril; Alatan, Abdullah Aydın; Department of Electrical and Electronics Engineering (2015) Deep learning is the discipline of training computational models that are composed of multiple layers and these methods have improved the state of the art in many areas such as visual object detection, scene understanding or speech recognition. Rebirth of these fairly old computational models is usually related to the availability of large datasets, increase in the computational power of current hardware and more recently proposed unsupervised training methods that exploit the internal structure of very lar...
3D object recognition from range images İzciler, Fatih; Halıcı, Uğur; Department of Electrical and Electronics Engineering (2012) Recognizing generic objects by single or multi view range images is a contemporary popular problem in 3D object recognition area with developing technology of scanning devices such as laser range scanners. This problem is vital to current and future vision systems performing shape based matching and classification of the objects in an arbitrary scene. Despite improvements on scanners, there are still imperfections on range scans such as holes or unconnected parts on images. This studyobjects at proposing an...

Citation Formats

E. Akdağ, “Human behavior understanding through 3D data,” M.S. - Master of Science, Middle East Technical University, 2015.