Please use this identifier to cite or link to this item:
Title: Motion and emotion : Semantic knowledge for hollywood film indexing
Keywords: affective motion classification film grammar shot
Issue Date: 2-May-2008
Citation: WANG HEE LIN (2008-05-02). Motion and emotion : Semantic knowledge for hollywood film indexing. ScholarBank@NUS Repository.
Abstract: The thesis proposes novel frameworks for the automated classification of affective and directing related higher-level semantic knowledge from Hollywood domain multimedia to augment indexing. We have proposed a framework grounded in psychology and cinematography to understand affective content at the scene level, as well as a comprehensive set of emotional categories for affective classification. The framework is validated by the largest and most broad-based video corpus in affective work to date. We investigated into the pivotal relationship between motion and higher level cinema semantics using the philosophy of film directing grammar. A novel MRF motion segmentation algorithm with the intrinsic ability to model simple occlusion has been proposed to enhance segmentation accuracy. The constraints inherent in directing grammar are exploited to construct a suitable and coherent directing semantics taxonomy. We demonstrate on Hollywood shots that motion alone can effectively index many types of the most significant directing semantic events.
Appears in Collections:Ph.D Theses (Open)

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
Wang Hee Lin Thesis_for upload.pdf3.82 MBAdobe PDF



Google ScholarTM


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.