Claude S. Fillion

Detecting content adaptive scaling of images for forensic applications

Content-aware resizing methods have recently been developed, among which, seam-carving has achieved the most widespread use. Seam-carvings versatility enables deliberate object removal and benign image resizing, in which perceptually important content is preserved. Both types of modifications compromise the utility and validity of the modified images as evidence in legal and journalistic applications. It is therefore desirable that image forensic techniques detect the presence of seam-carving. In this paper we address detection of seam-carving for forensic purposes. As in other forensic applications, we pose the problem of seam-carving detection as the problem of classifying a test image in either of two classes: a) seam-carved or b) non-seam-carved. We adopt a pattern recognition approach in which a set of features is extracted from the test image and then a Support Vector Machine based classifier, trained over a set of images, is utilized to estimate which of the two classes the test image lies in. Based on our study of the seam-carving algorithm, we propose a set of intuitively motivated features for the detection of seam-carving. Our methodology for detection of seam-carving is then evaluated over a test database of images. We demonstrate that the proposed method provides the capability for detecting seam-carving with high accuracy. For images which have been reduced 30% by benign seam-carving, our method provides a classification accuracy of 91%.

Image simulation for automatic license plate recognition

Automatic license plate recognition (ALPR) is an important capability for traffic surveillance applications, including toll monitoring and detection of different types of traffic violations. ALPR is a multi-stage process comprising plate localization, character segmentation, optical character recognition (OCR), and identification of originating jurisdiction (i.e. state or province). Training of an ALPR system for a new jurisdiction typically involves gathering vast amounts of license plate images and associated ground truth data, followed by iterative tuning and optimization of the ALPR algorithms. The substantial time and effort required to train and optimize the ALPR system can result in excessive operational cost and overhead. In this paper we propose a framework to create an artificial set of license plate images for accelerated training and optimization of ALPR algorithms. The framework comprises two steps: the synthesis of license plate images according to the design and layout for a jurisdiction of interest; and the modeling of imaging transformations and distortions typically encountered in the image capture process. Distortion parameters are estimated by measurements of real plate images. The simulation methodology is successfully demonstrated for training of OCR.

Adaptive removal of background and white space from document images using seam categorization

Document images are obtained regularly by rasterization of document content and as scans of printed documents. Resizing via background and white space removal is often desired for better consumption of these images, whether on displays or in print. While white space and background are easy to identify in images, existing methods such as naïve removal and content aware resizing (seam carving) each have limitations that can lead to undesirable artifacts, such as uneven spacing between lines of text or poor arrangement of content. An adaptive method based on image content is hence needed. In this paper we propose an adaptive method to intelligently remove white space and background content from document images. Document images are different from pictorial images in structure. They typically contain objects (text letters, pictures and graphics) separated by uniform background, which include both white paper space and other uniform color background. Pixels in uniform background regions are excellent candidates for deletion if resizing is required, as they introduce less change in document content and style, compared with deletion of object pixels. We propose a background deletion method that exploits both local and global context. The method aims to retain the document structural information and image quality.

The use of spatially based complexity measures towards color gamut mapping and image resizing

Several color-imaging algorithms such as color gamut mapping to a target device and resizing of color images have traditionally involved pixel-wise operations. That is, each color value is processed independent of its neighbors in the image. In recent years, applications such as spatial gamut mapping have demonstrated the virtues of incorporating spatial context into color processing tasks. In this paper, we investigate the use of locally based measures of image complexity such as the entropy to enhance the performance of two color imaging algorithms viz. spatial gamut mapping and content-aware resizing of color images. When applied to spatial gamut mapping (SGM), the use of these spatially based local complexity measures helps adaptively determine gamut mapping parameters as a function of image content - hence eliminating certain artifacts commonly encountered in SGM algorithms. Likewise, developing measures of complexity of color-content in a pixel neighborhood can help significantly enhance performance of content-aware resizing algorithms for color images. While the paper successfully employs intuitively based measures of image complexity, it also aims to bring to light potentially greater rewards that may be reaped should more formal measures of local complexity of color content be developed.