Range-Doppler estimation algorithms that use the phase of the returned signal face the 2π phase ambiguity problem. In this letter, we eliminate phase wrapping by using two coincident opposite polarity chirps but equal starting frequencies. The returns for both chirps from the same target have near equal phases. The phase wrap is present in both returns but they are the same. Therefore, the differential phase matching used to pair the returns from the same target cancels out the phase wrap. No phase unwrapping is needed. The proposed method has no inherent maximum unambiguous range or velocity, works in multitarget theaters, requires a much lower sampling rate, is not affected by range migration, decouples the Doppler cycle from the pulse repetition frequency and is computationally more efficient than FFT-based methods.

An analysis of the relationship between multipath ghosts and the direct target image for radar imaging is presented. A multipath point spread function (PSF) is defined that allows for specular reflections in the local environment and can allow the ghost images to be localized. Analysis of the multipath PSF shows that certain ghosts can only be focused for the far field synthetic aperture radar case and not the full array case. Importantly, the ghosts are shown to be equivalent to direct target images taken from different observation angles. This equivalence suggests that exploiting the ghosts would improve target classification performance, and this improvement is demonstrated using experimental data and a naïve Bayesian classifer. The maximum performance gain achieved is 32%.

A through-the-wall radar image (TWRI) bears little resemblance to the equivalent optical image, making it difficult to interpret. To maximize the intelligence that may be obtained, it is desirable to automate the classification of targets in the image to support human operators. This paper presents a technique for classifying stationary targets based on the high-range resolution profile (HRRP) extracted from 3-D TWRIs. The dependence of the image on the target location is discussed using a system point spread function (PSF) approach. It is shown that the position dependence will cause a classifier to fail, unless the image to be classified is aligned to a classifier-training location. A target image alignment technique based on deconvolution of the image with the system PSF is proposed. Comparison of the aligned target images with measured images shows the alignment process introducing normalized mean squared error (NMSE) ≤ 9%. The HRRP extracted from aligned target images are classified using a naive Bayesian classifier supported by principal component analysis. The classifier is tested using a real TWRI of canonical targets behind a concrete wall and shown to obtain correct classification rates ≥ 97%.

We propose an algorithm to embed data directly in the bitstream of JPEG imagery. The motivation for this approach is that images are seldom available in uncompressed form. Algorithms that operate in spatial domain, or even in coefficient domain, require full (or at best) partial decompression. Our approach exploits the fact that only a fraction of JPEG code space is actually used by available encoders. Data embedding is performed by mapping a used variable length code (VLC) to an unused VLC. However, standard viewers unaware of the change will not properly display the image. We address this problem by a novel error concealment technique. Concealment works by remapping run/size values of marked VLCs so that standard viewers do not lose synchronization and displays the image with minimum loss of quality. It is possible for the embedded image to be visually identical to the original even though the two files are bitwise different. The algorithm is fast and transparent and embedding is reversible and file-size preserving. Under certain circumstances, file size may actually decrease despite carrying a payload.

In this paper, we propose a known-host-state methodology for designing image watermarks that are particularly robust to compression. The proposed approach outperforms traditional spread spectrum watermarking across all JPEG quality factors. The fundamental approach uses 2D chirps as spreading functions, followed by a chirp transform, to recover the watermark. Because this method can spectrally shape the chirp to match image content and JPEG quantization, its performance is greatly enhanced. The energy localization of the chirp is exploited to embed low power watermark per image blocks while maintaining reliable detection performance.

In this letter, we propose a fragile digital watermarking algorithm that is directly applicable to compressed bit streams and is mathematically lossless with practically no increase in file size. The algorithm exploits the inefficiency of the code space of popular compression standards by recognizing that not all code words appear in the bit stream. By mapping the code words to the outside of the used code space, we can embed the watermark into the stream as forced bit errors. The decoder recognizes such code words, recovers the watermark from them, and restores the media to its original state. As proof of the concept, the algorithm is then applied to the JPEG bit stream.

A recursive algorithm is proposed for synthesizing a discrete-time periodic signal from a specified discrete-time discrete-frequency Wigner distribution by minimizing the error in the cyclic outer product. Equating the gradient of the error with respect to signal samples to zero results in a set of nonlinear simultaneous equations, which are solved using Newton-Raphson's method. A simulation example is given showing the synthesized signal over the first few iterations.

This paper reports on watermarking of MPEG video for authentication and tamper detection. The algorithm operates entirely on the compressed bitstream and maintains MPEG-2 syntax throughout. In the proposed algorithm, I-frames watermark the entire length of the following GOP. The algorithm first computes the payload capacity of each GOP by identifying the population of eligible VLCs capable of carrying the watermarks. The candidate I-frame is then compressed to this target bit rate and the bits are reordered based on a GOP-unique key. The scrambled bits then sequentially watermark the GOP. Detection follows an inverse process. Watermarked video is resistant to tampering at the GOP level by observing the key sequencing needed for watermark extraction. Missing GOPs cause jumps in the required keys which can then be connected to the number of affected GOPs. Corruption or frame-level tampering inside a GOP can also be spotted using a similar approach.

Filtering in time-frequency domain involves synthesis of signals with specific time-frequency signature. We observe that watermarking of time-frequency distributions is a similar problem. The proposed domain is the joint time-frequency cells of the Wigner distribution of the image. Watermarked cells are selected based on their impact on image quality as well as resilience to JPEG. Inversion of the watermarked Wigner is a signal synthesis problem and is inherently different from inverse DCT. Experimental results illustrate that time-frequency space has greater capacity and security advantage over spatial or spectral domains alone.

We report on the development of a watermarking algorithm designed for video authentication and tamper detection. The objectives are to determine unauthorized cut-and-splice or cut-insert-splice operation and quantify the extent of such editing. We demonstrate that the proposed algorithm can identify cut start and duration down to single frame precision. The approach embeds a watermark with a strong timing content, violation of which can be traced back to the parameters of the editing operation.

In this work we apply a direct sequence spread spectrum model to the watermarking of digital video. First, the video signal is modeled as a sequence of bit planes arranged along the time axis. Watermarking of this sequence is a two layer operation. A controlling m-sequence first establishes a pseudorandom order in the bitplane stream for later watermarking. Watermarks, defined as m-frames, supplant the tagged bitplanes. We have shown that the watermark, when limited to the 4 lowest bitplanes of an 8-bit video, are unnoticeable. Moreover, attempts in corrupting the image to destroy the watermark renders the video useless before damaging the seal itself. The watermarked video is also robust to video editing attempts such as subsampling, frame reordering etc. The watermark is also identifiable from very short segments of video. Individual frames extracted from the video will also contain copyright information.

Camera calibration is an important task in a computer vision system. This step establishes the all important relationship between the 2D image coordinates and their 3D counterparts and vice versa. Image center is subject to many different definitions. Moreover, its location is subject to a drift in most variable parameter lenses. This drift has an adverse effect on algorithms that rely on the a fixed center of image. A geometric model is presented that describes the trajectory of feature points in response to changing focal length under a shifting image center.

Optic flow has shown to be a rich source of 3-dimensional scene structure. In this work we report on an approach that induces optic flow without camera motion. This approach has been suggested before but the core issue, shape matching across focal lengths was not addressed. We propose a model-based matching technique based on the physics of optical zooming.>

Images taken with long telephoto lenses exhibit a characteristic perspective distortion known as compression of space. Objects that are actually far apart appear unusually close together, and observed texture gradients and optic flows impart a distorted sense of orientation and depth. An attempt is made to quantify the space compression. It is shown that only a full perspective model, rare in shape from texture literature for example, is able to compute the above effect.>

Complexity of hierarchical representation of images is defined as the total number of nodes in the representation tree. An a priori knowledge of this quantity is of considerable interest in problems involving tree search, storage and transmission of imagery. This paper proposes a systematic approach for the evaluation of quadtree complexity that is based on a flexible linkage paradigm. It is further realized that a quadtree may undergo a complexity reduction through node condensation. This event is fully modeled and absorbed in the expected complexity expression through a multidimensional weighting function. Inspection of the weighting surface provides a more clear view of the interaction of quadtree complexity and the random image model.>

The paper reports on the development of a machine vision system for assessing targeting accuracy of ballistic, projectile-firing weapon systems. Current techniques rely on either manual optical sighting or acoustic signature to locate the point of impact. Optical sighting, still the predominant method in many events, is manual and imprecise. Acoustic-based approaches automate the process but require multiple sensor placements. The machine vision system described is able to continuously monitor the target, report precise quantitative targeting information and simultaneously provide a color-coded display of impacts. Special provisions have been built-in to account for target plane motion and overlapping impacts phenomenon.>

The soft-linked quadtree is proposed as a hierarchical representation paradigm for random fields. It is shown that, through the use of such a model, a priori information can be obtained on the complexity of a quadtree. Because of this property, a parametric analysis can be substituted for simulation. However, this capability is only one of the many such features of a soft-linked model. Almost all the concepts defined over a conventional quadtree, e.g., neighbor and neighbor finding, perimeter and area, etc., can be revisited. Since the notions of state probability and link strength are means of expressing confidence in assigned labels, the soft-linked model can be further developed into a hierarchical representation paradigm for uncertainty.>

The quadtree is a variant of the maximal block representation scheme. The author introduces the concept of flexible linkage and a modular ring structure. A soft-linked quadtree can be put together by various admissable permutations of three rings. Each ring consists of two nodes with uniform and/or mixed label. Each node, in contrast to a conventional quadtree, has only a probabilistic affiliation to various segments of an image. Moreover, the links between father and son nodes are soft and characterized by a branch strength. All nodes have children; the degree of mutual attachment is the variable. The presence of soft links gives rise to an up projection effect whereby the classification of children impacts the ancestor's original label. This impact is modeled as a recursive relationship between all nodal state probabilities, branch strengths and tree depth. The implication of this approach for some existing quadtree-based algorithm is discussed.>