Full Record

New Search | Similar Records

Title Locally Tuned Nonlinear Manifold for Person Independent Head Pose Estimation
Publication Date
Degree PhD
Discipline/Department Electrical Engineering
Degree Level doctoral
University/Publisher University of Dayton
Abstract Fine-grain head pose estimation from imagery is an essential operation for many human-centered systems, including pose independent face recognition and human-computer interaction systems. It is only recently that estimation systems have evolved past a coarse level classification of pose and concentrated on fine-grain estimation. In particular, the state of the art of such systems consists of nonlinear manifold embedding techniques, such as Locally Linear Embedding and Isomap, that capture the intrinsic relationship of a pose varying face dataset. The success of these solutions can be attributed to the acknowledgment that image variation corresponding to pose change is nonlinear in nature. Yet, these algorithms are limited by the complexity of embedding functions that describe the relationship and provide no clear method for projecting novel data to the latent space. On the other hand, linear methods and nonlinear approximation techniques permit a simple embedding process, but lack the representational quality to globally describe the nonlinear image variation. In this dissertation, a pose estimation framework that seeks to describe the global nonlinear relationship in terms of localized linear functions is presented. A locally tuned nonlinear manifold is formulated using a two-layer system based on the assumptions that coarse pose estimation can be performed adequately using supervised linear methods, and fine pose estimation can be achieved using linear regressive functions if the scope of the pose manifold is limited. The localized linear approach results in a simplistic model for which probe input can be embedded through a cascade of linear transformations. Additionally, new methods for modeling the localized structures using feature enhanced Canonical Correlation Analysis are developed, where pose variation is regarded as a continuous variable and is represented by a manifold in feature space. The feature enhanced methods are used to identify the modes of correlation between the observed input images and the head pose angle. These techniques exploit oriented filters which serve two key purposes: (a) eliminate noise features, while boosting image elements that are associated with head pose (b) provide multiple dimensions of the input, allowing the correlation analysis process to extract more basis vectors to provide higher accuracy. A pose estimation system is implemented utilizing simple linear subspace methods, phase congruency, and Gabor features. The framework is tested using conventional test strategies and widely accepted pose-varying face databases. The proposed method is first tested using homogeneous datasets, where the training and testing face images are sampled from the same database. The generalization capabilities of the system are then tested using heterogeneous tests, where the training data is taken from a different database than the testing set. The proposed system is shown to perform fine head pose estimation with competitive accuracy when compared with state of the art nonlinear manifold…
Subjects/Keywords Computer Engineering; Electrical Engineering; Head Pose Estimation; Piecewise Linear Manifold; Pose Sensitive Representations; Coarse to Fine; Head Orientation; Phase Congruency
Contributors Asari, Vijayan (Committee Chair)
Language en
Rights unrestricted ; This thesis or dissertation is protected by copyright: all rights reserved. It may not be copied or redistributed beyond the terms of applicable copyright laws.
Country of Publication us
Format application/pdf
Record ID oai:etd.ohiolink.edu:dayton1311968002
Repository ohiolink
Date Retrieved
Date Indexed 2021-01-29
Grantor University of Dayton

Sample Search Hits | Sample Images

…with local linear functions to describe portions of the global pose manifold. These local pose models are supported by a supervised neighborhood estimation process. In effect, a coarse to fine pose estimation method is defined with the intent of…

…functions (Pφ ) are created to locally describe pose change. Fine pose estimation is performed by propagating through two layers of the model which include: Layer 1 - Coarse pose estimation is performed using class-based supervised linear…

…linear model of the pose manifold that is supported by coarse pose estimation. We hypothesize that by defining the system in a coarse to fine configuration, we can model the latent structure both simplistically and accurately. 1.1.2 Locally Tuned…

…experience occlusion. Appearance-based methods Class based pose estimation is performed using pattern classification methods similar to those used for face recognition. Coarse pose estimation can function well using these techniques. However, this class of…

to distinct pose angles. Detector arrays are computationally expensive and are unable to perform fine/continuous pose estimation. Dimensionality reduction and manifold learning A linear or nonlinear embedding process is used to describe the latent…

…data set. For each region, Rφ , projections are created to describe pose change (Pφ ). Layer 2 - Fine pose estimation is performed using region dependent pose-regressive transforms. It is clear from figure 1.2 that we have created a piecewise…

…Manifolds An integral portion of the framework, described in section 1.1.1, is the region dependent pose-regressive transforms, or locally tuned manifolds. These localized transforms serve to refine the coarse estimate performed by the 1st layer of the…

…training images to create a subspace that discriminates the global manifold neighborhoods. For each neighborhood φ, local linear transforms Pφ are created to model the fine pose variation of the localized dataset. Figure 1.4 illustrates the process of…