While image stabilization(IS ) has become a default functionality for most digital cameras, there is a lack of
automatic IS evaluation scheme. Most publicly known camera IS reviews either require human visual assessment
or resort to some generic blur metric. The former is slow and inconsistent, and the latter may not be easily
scalable with respect to resolution variation and exposure variation when comparing different cameras. We
proposed a histogram based automatic IS evaluation scheme, which employs a white noise pattern as shooting
target. It is able to produce accurate and consistent IS benchmarks in a very fast manner.
While previous work on lens identification by chromatic aberration succeeded in distinguishing lenses of different
model, the CA patterns obtained were not stable enough to support distinguishing different copies of the same
lens. This paper discusses on how to eliminate two major hurdles in the way of obtaining a stable lens CA pattern.
The first hurdle was overcome by using a white noise pattern as shooting target to supplant the conventional
but misalignment-prone checkerboard pattern. The second hurdle was removed by the introduction of the lens
focal distance, which had not received the attention it deserves. Consequently, we were able to obtain a stable
enough CA pattern distinguishing different copies of the same lens. Finally, with a complete view of the lens CA
pattern feature space, it is possible to fulfil lens identification among a large lens database.
Detection results obtained from an oracle can be used to reverse-engineer the underlying detector structure, or
parameters thereof. In particular, if a detector uses a common structure like correlation or normalized correlation,
detection results can be used to estimate feature space dimensionality, watermark strength, and detector threshold
values. Previous estimation techniques used a simplistic but tractable model for a watermarked image in the
detection cone of a normalized correlation detector; in particular a watermarked image is assumed to lie along the
axis of the detection cone, essentially corresponding to an image of zero magnitude. This produced useful results
for feature spaces of fewer dimensions, but increasingly imprecise estimates for larger feature spaces. In this paper
we model the watermarked image properly as a sum of a cover vector and approximately orthogonal watermark
vector, offsetting the image within the cone, which is the geometry of a detector using normalized correlation.
This symmetry breaking produces a far more complex model which boils down to a quartic equation. Although
it is infeasible to find its symbolic solution even with the aid of computer, our numerical analysis results show
certain critical behavior which reveals the relationship between the attacking noise strength and the detector
parameters. The critical behavior predicted by our model extends our reverse-engineering capability to the case of
detectors with large feature space dimensions, which is not uncommon in multimedia watermarking algorithms.
The square root law holds that acceptable embedding rate is sublinear in the cover size, specifically O(square root of n), in
order to prevent detection as the warden's data and thus detector power increases.
One way to transcend this law, at least in the i.i.d.case, is to restrict the cover to a chosen subset whose
distribution is close to that of altered data. Embedding is then performed on this subset; this replaces the
problem of finding a small enough subset to evade detection with the problem of finding a large enough subset
that possesses a desired type distribution.
We show that one can find such a subset of size asymptotically proportional to n rather than
the square root of n. This
works in the case of both replacement and tampering: Even if the distribution of tampered data depends on
the distribution of cover data, one can find a fixed point in the probability simplex such that cover data of that
distribution yields stego data of the same distribution.
While the transmission of a subset is not allowed, this is no impediment: wet paper codes can be used, else in
the worst case a maximal desirable subset can be computed from the cover by both sender and receiver without
communication of side information.
An emerging form of steganographic communication uses ciphertext to replace the output of a random or strong
pseudo-random number generator. PRNG-driven media, for example computer animated backdrops in video-conferencing
channels, can then be used as a covert channel, if the PRNG bits that generated a piece of content
can be estimated by the recipient.
However, all bits sent over such a channel must be computationally indistinguishable from i.i.d. coin flips. Ciphertext
messages and even key exchange datagrams are easily shaped to match this distribution; however, when
placing these messages into a continous stream of PRNG bits, the sender is unable to provide synchronization
markers, metadata, or error correction to ensure the message's location and proper decoding.
In this paper we explore methods for message transmission and steganographic key exchange in such a "coin
flip" channel. We establish that key exchange is generally not possible in this channel if an adversary possesses
even a modest noise budget. If the warden is not vigilant in adding noise, however, communication is very simple.
Inspired by results from the Break Our Watermarking System (BOWS) contest, we explored techniques to
reverse-engineer watermarking algorithms via oracle attacks. We exploit a principle called "superrobustness,"
which allows a watermarking algorithm to be characterized by its resistance to specific distortions. The generic
application of this principle to an oracle attack seeks to find a severe false alarm, or a point on the watermark
detection region as far as possible from the watermarked image.
For specific types of detection regions, these severe false positives can leak information about the feature
space as well as detector parameters. We explore the specific case of detectors using normalized correlation, or
From December 2005 to March of 2006, the Break Our Watermarking System (BOWS) contest challenged
researchers to break an image watermark of unknown design. The attacked images had to possess a minimum
quality level of 30 dB PSNR, and the winners would be those of highest average quality over three images.
Our research team won this challenge, employing the strategy of reverse-engineering the watermark before any
attempts to attack it in earnest. We determined the frequency transform, sub-band, and an exploitable quirk in
the detector that made it sensitive to noise spikes. Of interest is our overall methodology of reverse-engineering
through severe false alarms, and we introduce a new concept, "superrobustness," which despite its positive name
is a security flaw.