Searcharxiv⌕ Search

arXiv subjects

Cheng Chang

Publications and source records attributed to Cheng Chang.

46 records · Page 3Linked to original sources
arXiv2018

Real-time deep hair matting on mobile devices

Augmented reality is an emerging technology in many application domains. Among them is the beauty industry, where live virtual try-on of beauty products is of great importance. In this paper, we address the problem of live hair color augmentation. To achieve this goal, hair needs to be segmented quickly and accurately. We show how a modified MobileNet CNN architecture can be used to segment the hair in real-time. Instead of training this network using large amounts of accurate segmentation data, which is difficult to obtain, we use crowd sourced hair segmentation data. While such data is much simpler to obtain, the segmentations there are noisy and coarse. Despite this, we show how our system can produce accurate and fine-detailed hair mattes, while running at over 30 fps on an iPad Pro tablet.

cs.CV↗
arXiv2014

Mechanotunable monatomic metal structures at graphene edges

Monatomic metal (e.g. silver) structures could form preferably at graphene edges. We explore their structural and electronic properties by performing density functional theory based first-principles calculations. The results show that cohesion between metal atoms, as well as electronic coupling between metal atoms and graphene edges offer remarkable structural stability of the hybrid. We find that the outstanding mechanical properties of graphene allow tunable properties of the metal monatomic structures by straining the structure. The concept is extended to metal rings and helices that form at open ends of carbon nanotubes and edges of twisted graphene ribbons. These findings demostrate the role of graphene edges as an efficient one-dimensional template for low-dimensional metal structures that are mechanotunable.

cond-mat.mes-hall↗
arXiv2009

Interference channel capacity region for randomized fixed-composition codes

The randomized fixe-composition with optimal decoding error exponents are studied \cite{Raul_ISIT,Raul_journal} for the finite alphabet interference channel (IFC) with two transmitter-receiver pairs. In this paper we investigate the capacity region of the randomized fixed-composition coding scheme. A complete characterization of the capacity region of the said coding scheme is given. The inner bound is derived by showing the existence of a positive error exponent within the capacity region. A simple universal decoding rule is given. The tight outer bound is derived by extending a technique first developed in \cite{Dueck_RC} for single input output channels to interference channels. It is shown that even with a sophisticated time-sharing scheme among randomized fixed-composition codes, the capacity region of the randomized fixed-composition coding is not bigger than the known Han-Kobayashi \cite{Han_Kobayashi} capacity region. This suggests that the average behavior of random codes are not sufficient to get new capacity regions.

cs.IT↗
arXiv2009

On the rate distortion function of Bernoulli Gaussian sequences

In this paper, we study the rate distortion function of the i.i.d sequence of multiplications of a Bernoulli $p$ random variable and a gaussian random variable $\sim N(0,1)$. We use a new technique in the derivation of the lower bound in which we establish the duality between channel coding and lossy source coding in the strong sense. We improve the lower bound on the rate distortion function over the best known lower bound by $p\log_2\frac{1}{p}$ if distortion $D$ is small. This has some interesting implications on sparse signals where $p$ is small since the known gap between the lower and upper bound is $H(p)$. This improvement in the lower bound shows that the lower and upper bounds are almost identical for sparse signals with small distortion because $\lim\limits_{p\to 0}\frac{p\log_2\frac{1}{p}}{H(p)}=1$.

cs.IT↗
arXiv2009

Joint source-channel with side information coding error exponents

In this paper, we study the upper and the lower bounds on the joint source-channel coding error exponent with decoder side-information. The results in the paper are non-trivial extensions of the Csiszar's classical paper [5]. Unlike the joint source-channel coding result in [5], it is not obvious whether the lower bound and the upper bound are equivalent even if the channel coding error exponent is known. For a class of channels, including the symmetric channels, we apply a game-theoretic result to establish the existence of a saddle point and hence prove that the lower and upper bounds are the same if the channel coding error exponent is known. More interestingly, we show that encoder side-information does not increase the error exponents in this case.

cs.IT↗
arXiv2007

The source coding game with a cheating switcher

Motivated by the lossy compression of an active-vision video stream, we consider the problem of finding the rate-distortion function of an arbitrarily varying source (AVS) composed of a finite number of subsources with known distributions. Berger's paper `The Source Coding Game', \emph{IEEE Trans. Inform. Theory}, 1971, solves this problem under the condition that the adversary is allowed only strictly causal access to the subsource realizations. We consider the case when the adversary has access to the subsource realizations non-causally. Using the type-covering lemma, this new rate-distortion function is determined to be the maximum of the IID rate-distortion function over a set of source distributions attainable by the adversary. We then extend the results to allow for partial or noisy observations of subsource realizations. We further explore the model by attempting to find the rate-distortion function when the adversary is actually helpful. Finally, a bound is developed on the uniform continuity of the IID rate-distortion function for finite-alphabet sources. The bound is used to give a sufficient number of distributions that need to be sampled to compute the rate-distortion function of an AVS to within a certain accuracy. The bound is also used to give a rate of convergence for the estimate of the rate-distortion function for an unknown IID finite-alphabet source .

cs.IT↗
arXiv2007

The price of ignorance: The impact of side-information on delay for lossless source-coding

Inspired by the context of compressing encrypted sources, this paper considers the general tradeoff between rate, end-to-end delay, and probability of error for lossless source coding with side-information. The notion of end-to-end delay is made precise by considering a sequential setting in which source symbols are revealed in real time and need to be reconstructed at the decoder within a certain fixed latency requirement. Upper bounds are derived on the reliability functions with delay when side-information is known only to the decoder as well as when it is also known at the encoder. When the encoder is not ignorant of the side-information (including the trivial case when there is no side-information), it is possible to have substantially better tradeoffs between delay and probability of error at all rates. This shows that there is a fundamental price of ignorance in terms of end-to-end delay when the encoder is not aware of the side information. This effect is not visible if only fixed-block-length codes are considered. In this way, side-information in source-coding plays a role analogous to that of feedback in channel coding. While the theorems in this paper are asymptotic in terms of long delays and low probabilities of error, an example is used to show that the qualitative effects described here are significant even at short and moderate delays.

cs.IT↗
arXiv2007

The source coding game with a cheating switcher

Berger's paper `The Source Coding Game', IEEE Trans. Inform. Theory, 1971, considers the problem of finding the rate-distortion function for an adversarial source comprised of multiple known IID sources. The adversary, called the `switcher', was allowed only causal access to the source realizations and the rate-distortion function was obtained through the use of a type covering lemma. In this paper, the rate-distortion function of the adversarial source is described, under the assumption that the switcher has non-causal access to all source realizations. The proof utilizes the type covering lemma and simple conditional, random `switching' rules. The rate-distortion function is once again the maximization of the R(D) function for a region of attainable IID distributions.

cs.IT↗
arXiv2006

Lossless coding for distributed streaming sources

Distributed source coding is traditionally viewed in the block coding context -- all the source symbols are known in advance at the encoders. This paper instead considers a streaming setting in which iid source symbol pairs are revealed to the separate encoders in real time and need to be reconstructed at the decoder with some tolerable end-to-end delay using finite rate noiseless channels. A sequential random binning argument is used to derive a lower bound on the error exponent with delay and show that both ML decoding and universal decoding achieve the same positive error exponents inside the traditional Slepian-Wolf rate region. The error events are different from the block-coding error events and give rise to slightly different exponents. Because the sequential random binning scheme is also universal over delays, the resulting code eventually reconstructs every source symbol correctly with probability 1.

cs.IT↗