A hybrid feature model for seam carving detection

Document Type

Conference Proceeding

Publication Date

1-1-2017

Abstract

Seam carving, as a content-aware image resizing algorithm, is widely used nowadays. In this paper, an advanced hybrid feature model is presented to reveal the trace of seam carving, especially seam carving at a low carving rate, applied to uncompressed digital images. Two groups of features are designed to capture energy variation and pixel variation caused by seam carving, respectively. As indicated by the experimental works, the state-of-the-art performance on detecting 5% and 10% carving rate cases has been improved from 81.13% and 90.26% to 85.75% and 94.87%, respectively.

Identifier

85028464730 (Scopus)

ISBN

[9783319641843]

Publication Title

Lecture Notes in Computer Science Including Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics

External Full Text Location

https://doi.org/10.1007/978-3-319-64185-0_7

e-ISSN

16113349

ISSN

03029743

First Page

77

Last Page

89

Volume

10431 LNCS

This document is currently not available here.

Share

COinS