Skip to Main content Skip to Navigation
Conference papers

Semi-supervised Learning Based Aesthetic Classifier for Short Animations Embedded in Web Pages

Abstract : We propose a semi-supervised learning based computational model for aesthetic classification of short animation videos, which are nowadays part of many web pages. The proposed model is expected to be useful in developing an overall aesthetic model of web pages, leading to better evaluation of web page usability. We identified two feature sets describing aesthetics of an animated video. Based on the feature sets, we developed a Naïve-Bayes classifier by applying Co-training, a semi-supervised machine learning technique. The model classifies the videos as good, average or bad in terms of their aesthetic quality. We designed 18 videos and got those rated by 17 participants for use as the initial training set. Another set of 24 videos were designed and labeled using Co-training. We conducted an empirical study with 16 videos and 23 participants to ascertain the efficacy of the proposed model. The study results show 75% model accuracy.
Document type :
Conference papers
Complete list of metadata

Cited literature [31 references]  Display  Hide  Download
Contributor : Hal Ifip Connect in order to contact the contributor
Submitted on : Tuesday, March 28, 2017 - 4:38:08 PM
Last modification on : Tuesday, March 28, 2017 - 5:12:55 PM
Long-term archiving on: : Thursday, June 29, 2017 - 5:42:12 PM


Files produced by the author(s)


Distributed under a Creative Commons Attribution 4.0 International License



Dipak Bansal, Samit Bhattacharya. Semi-supervised Learning Based Aesthetic Classifier for Short Animations Embedded in Web Pages. 14th International Conference on Human-Computer Interaction (INTERACT), Sep 2013, Cape Town, South Africa. pp.728-745, ⟨10.1007/978-3-642-40483-2_51⟩. ⟨hal-01497475⟩



Record views


Files downloads