Non-rigid point registration is the process of aligning one set of points to another that is deformed from the former. It plays a key role in many computer vision, medical imaging and pattern recognition applications, such as generating cartoon animation, object retrieval, recovering dynamic motions of human organs and muscles, and template matching for hand-written characters. In this book, we introduce a novel approach for muscle function estimation using non-rigid point set registration.
The present study investigated whether the teaching of L2 vocabulary with still cartoon pictures and animated cartoon pictures would result in a significant difference in second language learners receptive and productive knowledge of the target words. Also, the effect of test type (receptive vs. productive) on participants retention of the target words was examined. Finally, the study tried to find out whether the semantic category of vocabulary words influences the rate of successful vocabulary retention across picture types. For the purpose, a group of 17 ESL students from a Midwestern University participated in both treatments with still and animated pictures, followed by vocabulary tests. The results showed that over 80% of the target words were successfully retrieved on the receptive knowledge tests vs. only about 40% successful retrieval on the productive knowledge tests. Yet, the results did not reveal significant differences in vocabulary gain due to picture type as both treatments showed similar success rate of retention of the target words, especially in view of receptive knowledge.
Sports video analysis and reconstruction, especially soccer, has received much of the attention due to its popularity as well as potential commercial value. A broad range of applications such as: content retrieval and indexing, semantic event highlight and summarization or objects recognition and tracking make building sports video analysis system a hot research area. By knowing the positions and movement trajectories of the players on the field, professionals could detect semantic events, analyze the teams tactics, strength and weakness as well as produce meaningful highlights and summarizations. The analysis of soccer video is complex due to its nature of congested, rapidly changing, and non-linear fashion of each player's movements. In this work, we present a soccer video analysis system for the Video-to-cartoon applications. The system is capable of performing detection from single and multiple camera views with robustness and efficiency, as well as generates the 3D virtual reality of the scene so users can interactively change the viewpoints while watching the match. Novel algorithms and techniques for building the system will be discussed in detail.
Deeper understanding on human actions is required in many applications, e.g., action recognition (security), animation (sport, 3D cartoon movies and virtual world), etc. The task of automatic extraction of semantic action features within the data set is gaining in importance. We propose novel approaches to automatically extract the action features from 3D Motion Capture data for high accuracy in real-time action recognition. In particular, we contribute to two different areas dealing with variation at different features levels, i.e., 1) Extract of Discriminate Patterns from Skeleton Sequences approach is based on features that are close to the raw data. It provides a foundation in lower dimensional representation for the movement sequence analysis, retrieval, identification and synthesis, and 2) Automatic Extraction of Semantic Action Features approach focuses on solving the high-dimensional computational problems arising from the human motion sequences. It supports the follow-up stages of processing the human movement on a natural language level. As one common underlying concept, the propose approach contains a retrieval component for extracted action features.
Cartoon images play essential roles in our everyday lives especially in entertainment, education, and advertisement that become an increasingly intensive research in the field of multimedia and computer graphics. Cartoon image retrieval is bedrock of producer's new cartoon videos, particularly when provides initial planning to be elevated in so as to identify their intention of artistic. Today, a number of researchers have exploited the concepts related to content based images retrieval (CBIR) to search for cartoon images containing particular object(s) of interest. This work aims to design content-based retrieval system for cartoon images. The proposed images retrieval system consists of four basic stages "(image quantization, object extraction, feature extraction and similarity matching). The system extracts features for the query image which is a percentage of each index value (dominant color), cluster values for three bands of quantization image (RGB), and the weight of each index values depends on its location in the image, then compares it with the images features in database system in similarity matching stage and then the top 10 images are retrieved as results.