Spectral methods to determine the exact scaling factor of resampled digital images

Detecting Trace of Seam Carving for Forensic Analysis

“…Feng et al extracted 19-features from the normalized energy density of second derivative images in the frequency domain, and then the feature vectors were applied to train and test a support vector machine (SVM) of resizing process [9]. To determine whether an investigated image is upscaled or downscaled, Pfennig and Kirchner combined Feng et al's approach with Popescu's linear predictor [10]. However, to the best of our knowledge, all the detectors still have difficulties in detecting the trace of content-aware image resizing, which intentionally carves out or stretches the part of the image without periodic resampling process.…”

“…Therefore, such artifacts may indicate a manipulation while examining the statistical properties of an image under investigation. For instance, many research works have revealed these inherent artifacts against manipulation: resampling [4]- [10], partial manipulation [11], [12], double quantization [13], [14], contrast enhancement [15], or sharpening [16].…”

“…To detect the resizing artifacts, several detectors exploiting the periodic- ity of interpolation through a resized image have been researched [4]- [10]. However, these detectors assume that the entire image is altered by a periodic resampling procedure.…”

Replacing DWT with DTCWT in blind image rotation angle estimation

“…For example, an imported image patch can be rotated within the host image in order to mimic a genuine photographed object. Hence, revealing a localized geometric transformation and possibly further estimating its exact parameters can provide meaningful clues for forensic analysis [1][2][3][4][5]. In this paper, we focus on the rotation angle estimation which is typically involved in image splicing, and can also occur in in-camera video stabilization [6,7].…”

Linear Filter Kernel Estimation Based on Digital Camera Sensor Noise

“…Even with practicable solutions to overcome issues of scalability in multi-classification [4,5,8], it seems infeasible to assume that a single model will ever be informative enough to discern between a large number of parameter settings across multiple processing types. Targeted techniques with the ability to estimate parameters of specific processing operations, such as previous JPEG quantization tables [9,10], the shape of non-linear intensity mappings [11], or parameters of affine transformations [12][13][14] will thus remain to play an important role in the realm of image forensics.…”