Thesis (MScIng)--University of Stellenbosch, 2005. / Bibliography. Also available via the Internet.

Thesis (Ph. D.)--Oregon Graduate Center, 1986.

Thesis (Ph. D.)--University of Washington, 2000. / Vita. Includes bibliographical references (leaves 84-92).

Hand-printed Chinese character recognition presents an interesting problem for Artificial Intelligence research. Input data in the form of arrays of pixel values cannot be directly mapped to unique character identifications because of the complexity of the characters. Thus, intermediate data structures are necessary, which in turn lead to a need to represent knowledge of the characters' composition. Building the intermediate constructs for these hand-printed characters necessarily involves choices among ambiguities, the set of which is so large that an efficient search algorithm becomes central to the recognition process. Schema labelling is a theory of how knowledge should be organized for recognition tasks in which composition structure is inherent in the domain, the composition entails ambiguity, and the ambiguity generates large search spaces. This thesis describes an implementation of an enhanced version of schema labelling for Chinese characters. The specific problems addressed by the enhancements, with some success, are (i) the segmentation of real images into objects usable by the schema system, (ii) the definition of schemas which adequately describe the generic composition of hand-printed Chinese characters, as well as common variations or vagaries, and (iii) the inclusion of sufficient "control knowledge" to prevent combinatorial explosion of the backtracking recognition process. Test characters for recognition systems can be classified along several dimensions. On the spectrum from type-set, through hand-printed, to hand-written forms, our system was tested on restricted hand-print, at a level somewhat more difficult than is normally attempted. On the spectrum of input types, from grey-scale pixel input through on-line stroke representations, our system was fully tested only at the high end, with complete synthetic strokes. We obtained a success rate of 57%, 12 out of the 21 characters tested. The principal success of the work is that characters of the complexity tested could be recognized at all, and in the impact schema labelling techniques had on that recognition. / Science, Faculty of / Computer Science, Department of / Graduate

No description available.

No description available.

No description available.

Wong Chun Nang. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2001. / Includes bibliographical references (leaves 56-58). / Abstracts in English and Chinese.

For face identification, especially by human, it is desirable to render a high-resolution (HR) face image from the low-resolution (LR) one, which is called face hallucination or face super-resolution. A number of super-resolution techniques have been proposed in recent years. However, for face hallucination the utilization of the special properties of the faces is conductive to generate the HR face images. / In this thesis, we propose a new face hallucination framework based on image patches, which integrates two novel statistical super-resolution models. Considering that image patches reflect the combined effect of personal characteristics and patch-location, we first formulate a TensorPatch model based on multilinear analysis to explicitly model the interaction between multiple constituent factors. Motivated by Locally Linear Embedding, we develop an enhanced multilinear patch hallucination algorithm, which efficiently exploits the local distribution structure in the sample space. To better preserve face subtle details, we derive the Coupled PCA algorithm to learn the relation between HR residue and LR residue, which is utilized for compensate the error residue in hallucinated images. Experiments demonstrate that our framework not only well maintains the global facial structures, but also recovers the detailed facial traits in high quality. (Abstract shortened by UMI.) / In this thesis, we propose a novel dimensionality reduction algorithm called graph-regularized projection (GRP) to tackle the problem of semi-supervised dimensionality reduction that is rarely investigated in the literature. Given partially labeled data points, GRP aims at learning a not only smooth but also discriminative projection from high-dimensional data vectors to their latent low-dimensional representations. Motivated by recent semi-supervised learning process: graph regularization, we develop a graph-based regularization framework to enforce smoothness along the graph of the desired projection initiated by margin maximization. As a result, GRP has a natural out-of-sample extension to novel examples and thus can be generalized to the entire high-dimensional space. Extensive experiments on a synthetic dataset and several real databases demonstrate the effectiveness of our algorithm. / Next, this thesis addresses the problem of how to learn an appropriate feature representation from video to benefit video-based face recognition. By simultaneously exploiting the spatial and temporal information, the problem is posed as learning Spatio-Temporal Embedding (STE) from raw videos. STE of a video sequence is defined as its condensed version capturing the essence of space-time characteristics of the video. Relying on the co-occurrence statistics and supervised signatures provided by training videos, STE preserves the intrinsic temporal structures hidden in video volume, meanwhile encodes the discriminative cues into the spatial domain. To conduct STE, we propose two novel techniques, Bayesian keyframe learning and nonparametric discriminant embedding (NDE), for temporal and spatial learning, respectively. In terms of learned STEs, we derive a statistical formulation to the recognition problem with a probabilistic fusion model. On a large face video database containing more than 200 training and testing sequences, our approach consistently outperforms the state-of-the-art methods, achieving a perfect recognition accuracy. / Liu, Wei. / "August 2007." / Advisers: Xiaoou Tang; Jianzhuang Liu. / Source: Dissertation Abstracts International, Volume: 69-02, Section: B, page: 1110. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2007. / Includes bibliographical references (p. 140-151). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Electronic reproduction. [Ann Arbor, MI] : ProQuest Information and Learning, [200-] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract in English and Chinese. / School code: 1307.

Wang, Yingze. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2009. / Includes bibliographical references (leaves 42-47). / Abstract also in Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 2 --- Related Work --- p.5 / Chapter 2.1 --- Line labeling and realization problem --- p.5 / Chapter 2.2 --- 3D reconstruction from multiple views --- p.6 / Chapter 2.3 --- 3D reconstruction from single line drawings --- p.7 / Chapter 2.3.1 --- Face identification from the line drawings --- p.7 / Chapter 2.3.2 --- 3D geometry reconstruction --- p.9 / Chapter 2.4 --- Our research topic and contributions --- p.13 / Chapter 3 --- Reconstruction of Curved Manifold Objects --- p.14 / Chapter 3.1 --- Assumptions and terminology --- p.14 / Chapter 3.2 --- Reconstruction of curved manifold objects --- p.17 / Chapter 3.2.1 --- Distinguishing between curved and planar faces --- p.17 / Chapter 3.2.2 --- Transformation of Line Drawings --- p.20 / Chapter 3.2.3 --- Regularities --- p.23 / Chapter 3.2.4 --- 3D Wireframe Reconstruction --- p.26 / Chapter 3.2.5 --- Generating Curved Faces --- p.28 / Chapter 3.2.6 --- The Complete 3D Reconstruction Algorithm --- p.33 / Chapter 4 --- Experiments --- p.35 / Chapter 5 --- Conclusions and Future Work --- p.40 / Chapter 5.1 --- Conclusions --- p.40 / Chapter 5.2 --- Future work --- p.40 / Bibliography --- p.42