Semi-automatic video object segmentation using LVQ with color and spatial features

被引：3

作者：

Mochamad, H ^{[1
]}

Loy, HC ^{[1
]}

Aoki, T ^{[1
]}

机构：

[1] Tohoku Univ, Grad Sch Informat Sci, Sendai, Miyagi 9808579, Japan

来源：

IEICE TRANSACTIONS ON INFORMATION AND SYSTEMS | 2005年 / E88D卷 / 07期

关键词：

object extraction; object segmentation; learning vector quantization; video processing; MPEG;

D O I：

10.1093/ietisy/e88-d.7.1553

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

This paper presents a semi-automatic algorithm for video object segmentation. Our algorithm assumes the use of multiple key video frames in which a semantic object of interest is defined in advance with human assistance. For video frames between every two key frames, the specified video objectis tracked and segmented automatically using Learning Vector Quantization (LVQ). Each pixel of a video frame is represented by a 5-dimensional feature vector integrating spatial and color information. We introduce a parameter K to adjust the balance of spatial and color information. Experimental results demonstrate that the algorithm can segment the video object consistently with less than 2% average error when the object is moving at a moderate speed.

引用

页码：1553 / 1560

页数：8