Detecting Trace of Seam Carving for Forensic Analysis

Seung-Jin RYU  Hae-Yeoun LEE  Heung-Kyu LEE  

IEICE TRANSACTIONS on Information and Systems   Vol.E97-D   No.5   pp.1304-1311
Publication Date: 2014/05/01
Online ISSN: 1745-1361
DOI: 10.1587/transinf.E97.D.1304
Type of Manuscript: PAPER
Category: Image Recognition, Computer Vision
image forensics,  content-aware image resizing,  seam carving,  seam insertion,  

Full Text: PDF(2.3MB)>>
Buy this Article

Seam carving, which preserves semantically important image content during resizing process, has been actively researched in recent years. This paper proposes a novel forensic technique to detect the trace of seam carving. We exploit the energy bias and noise level of images under analysis to reliably unveil the evidence of seam carving. Furthermore, we design a detector investigating the relationship among neighboring pixels to estimate the inserted seams. Experimental results from a large set of test images indicates the superior performance of the proposed methods for both seam carving and seam insertion.