We consider asynchronous communication over point-to-point discrete memoryless channels without feedback. The transmitter starts sending one block codeword at an instant that is uniformly distributed within a certain time period, which represents the level of asynchronism. The receiver, by means of a sequential decoder, must isolate the message without knowing when the codeword transmission starts but being cognizant of the asynchronism level. We are interested in how quickly can the receiver isolate the sent message, particularly in the regime where the asynchronism level is exponentially larger than the codeword length, which we refer to as 'strong asynchronism. ' This model of sparse communication might represent the situation of a sensor that remains idle most of the time and, only occasionally, transmits information to a remote base station which needs to quickly take action. Because of the limited amount of energy the sensor possesses, assuming the same cost per transmitted symbol, it is of interest to consider minimum size codewords given the asynchronism level.The first result is an asymptotic characterization of the largest asynchronism level, in terms of the codeword length, for which reliable communication can be achieved: vanishing error probability can be guaranteed as the codeword length N tends to infinity while the asynchronism level grows as e N α if and only if α does not exceed the synchronization threshold, a constant that admits a simple closed form expression, and is at least as large as the capacity of the synchronized channel.The second result is the characterization of a set of achievable strictly positive rates in the regime where the asynchronism level is exponential in the codeword length, and where the rate is defined with respect to the expected (random) delay between the time information starts being emitted until the time the receiver makes a decision. Interestingly, this achievability result is obtained by a coding strategy whose decoder not only operates in an asynchronously, but has an almost universal decision rule, in the sense that it is almost independent of the channel statistics.As an application of the first result we consider antipodal signaling over a Gaussian additive channel and derive a simple necessary condition between blocklength, asynchronism level, and SNR for achieving reliable communication.

We consider the 'one-shot frame synchronization problem' where a decoder wants to locate a sync pattern at the output of a channel on the basis of sequential observations. We assume that the sync pattern of length N starts being emitted at a random time within some interval of size A, that characterizes the asynchronism level between the transmitter and the receiver. We show that a sequential decoder can optimally locate the sync pattern, i.e., exactly, without delay, and with probability approaching one as N → ∞, if and only if the asynchronism level grows as O(e N α ), with α below the synchronization threshold, a constant that admits a simple expression depending on the channel. This constant is the same as the one that characterizes the limit for reliable asynchronous communication, as was recently reported by the authors. If α exceeds the synchronization threshold, any decoder, sequential or non-sequential, locates the sync pattern with an error that tends to one as N → ∞. Hence, a sequential decoder can locate a sync pattern as well as the (non-sequential) maximum likelihood decoder that operates on the basis of output sequences of maximum length A + N − 1, but with much fewer observations.

Having accurate left ventricle (LV) segmentations across a cardiac cycle provides useful quantitative (e.g. ejection fraction) and qualitative information for diagnosis of certain heart conditions. Existing LV segmentation techniques are founded mostly upon algorithms for segmenting static images. In order to exploit the dynamic structure of the heart in a principled manner, we approach the problem of LV segmentation as a recursive estimation problem. In our framework, LV boundaries constitute the dynamic system state to be estimated, and a sequence of observed cardiac images constitute the data. By formulating the problem as one of state estimation, the segmentation at each particular time is based not only on the data observed at that instant, but also on predictions based on past segmentations. This requires a dynamical system model of the LV, which we propose to learn from training data through an information-theoretic approach. To incorporate the learned dynamic model into our segmentation framework and obtain predictions, we use ideas from particle filtering. Our framework uses a curve evolution method to combine such predictions with the observed images to estimate the LV boundaries at each time. We demonstrate the effectiveness of the proposed approach on a large set of cardiac images. We observe that our approach provides more accurate segmentations than those from static image segmentation techniques, especially when the observed data are of limited quality.

No abstract