@inproceedings{104be9c619f044ac8456d38223ea948a,
title = "Automatically detecting protruding objects when shooting environmental portraits",
abstract = "This study proposes techniques for detecting unintentional protruding objects from a subject's head in environmental portraits. The protruding objects are determined based on the color and edge information of the background regions adjacent to the head regions in an image sequence. The proposed algorithm consists of watershed segmentation and KLT feature tracking model for extracting foreground regions, a ROI (Region of Interest) extracting model based on face detection results, and a protruding object detection model based on the color clusters and edges of the background regions inside the ROI. Experimental evaluations using four test videos with different backgrounds, lighting conditions, and head ornaments show that the average detection rate and false detection rate of the proposed algorithm are 87.40% and 12.11% respectively.",
keywords = "Computational Photography, Photo Composition, Protruding Object",
author = "Lo, {Pei Yu} and Shih, {Sheng Wen} and Liu, {Jen Chang} and Hong, {Jen Shin}",
year = "2011",
doi = "10.1007/978-3-642-22819-3_14",
language = "English",
isbn = "9783642228186",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
number = "PART 2",
pages = "132--141",
booktitle = "Computer Vision - ACCV 2010 Workshops - ACCV 2010 International Workshops, Revised Selected Papers",
edition = "PART 2",
note = "International Workshops on Computer Vision, ACCV 2010 ; Conference date: 08-11-2010 Through 09-11-2010",
}