This work proposes a novel image restoration technique inspired by the Ulyanov’s deep image prior (DIP) method. DIP uses a deep convolutional network as an image prior to generate a restored image from a random input one, which brings an advantage of no requirement of training data. However, one problem arises that the interpretability is low and it is not trivial to explain the need for the random input. This paper contributes to train a network by using the Stein’s unbiased risk estimator (SURE) with Monte-Carlo computation for self-supervised learning of image restoration. The Monte-Carlo calculation accepts a random input for computing divergence of the network, and allows us to interpret the reason why a random input is needed. Our framework trains an image restorer instead of an image generator, which simplifies the interpretation of the network structure. Thanks to this, the Luisier’s interscale linear expansion of the thresholding (LET) can become introduced to exploit the cross-correlation among extracted features. In order to avoid using group delay compensation and reduce the number of design parameters, a structured DIP is constructed by using non-separable oversampled lapped transform (NSOLT) instead of U-Net. By showing some results of denoising simulation, the significance of the proposed method is verified.

This study extends a self-supervised image denoising technique proposed by the authors to a more general image restoration method. The previous work was inspired by Ulyanov’s deep image prior (DIP) method. DIP uses a deep convolutional network as an image prior to generate a restored image from a random one, which brings an advantage of no training data requirement. However, one problem arises that the interpretability is low and it is not trivial to explain the need for the random input. In the previous work, it is shown that Stein’s unbiased risk estimator (SURE) with Monte-Carlo computation can be used to train a denoising network. This approach allows us to interpret the random input requirement. The framework trains an image denoiser instead of an image generator. This work extends the discussion to more general image restoration problem by introducing a deep unrolling approach to reflect the measurement process. The new framework uses the Luisier’s inter-scale linear expansion of the thresholding (LET) for exploiting the cross-correlation among extracted features. Some simulation results of image restoration show the significance of the proposed method.

A new scheduling algorithm called serialised batch scheduling (SBS) for optical burst switching (OBS) networks is proposed. SBS aggregates and serialises bursts along a shared path into one composite burst, which is switched as one unit. There are two major processes in SBS, namely, batching and serialising. While the batching process chooses a set of bursts to form the composite burst, the serialising process determines how to organise the OBS bursts within the composite burst and generates a corresponding control packet for this burst. Several SBS batching and serialising schemes are introduced and analysed here. The study by the authors shows that the guard band and burst overlap can be reduced in the SBS and, thus, the packet loss rate and the number of switch reconfigurations can be significantly reduced. In addition, it is indicated that the proposed SBS algorithm can be coupled with other OBS scheduling algorithms and reduce the requirements for a high-speed optical switch in OBS networks.

In this paper, a new scheduling algorithm, Serialized Batch Scheduling (SBS) for Optical Burst Switching (OBS) networks is proposed. SBS aggregates and serializes bursts along shared path into one composite burst which is switched as a single unit. There are two important processes in SBS namely, batching and serializing. While the batching process chooses proper set of bursts to form the composite burst, the serializing process determines how to organize the OBS bursts within the composite burst and generates a corresponding composite control packet. Several SBS batching and serializing schemes are introduced and analyzed. Our study shows that the guard band and bursts overlap can be reduced in the proposed SBS, and therefore, the packet loss rate and the number of switch reconfigurations can be significantly reduced.

With the huge capacity, wavelength division multiplexing (WDM) optical networks are predominantly used as the transport infrastructure to carry inter-domain traffic for client networks. Traffic grooming refers to the aggregation of low-speed client traffic flows onto high-capacity optical connections, to achieve cost-effective traffic transport. In this paper, we propose a route segment technique to efficiently compute the client traffic blocking probability in dynamic traffic grooming, where client traffic randomly arrives/departs. In our experiments, the blocking probabilities computed by this technique closely match the results obtained by simulations.

Traffic grooming in wavelength division multiplexing (WDM) optical networks controls how to consolidate client calls with sub-wavelength data rates onto lightpaths. It can be classified into static or dynamic traffic grooming depending on whether the client traffic is static or dynamic. The principal problem in traffic grooming is to construct a logical topology to route client traffic over a given physical topology. For dynamic traffic grooming, the logical topology may be dynamically configured, or designed a priori given the stationary traffic demands between client nodes (and reconfigured on relatively large time scale, e.g., on the order of hours, to adapt to traffic demands changes). Both approaches have their pros and cons. This paper studies the latter case and develops a heuristic algorithm to design logical topology, constrained by client traffic blocking probability requirement and the maximum by-pass traffic amount allowed at each client node. We have compared logical topologies designed the heuristic and an ILP model. The heuristic performance is impressive.

Waveband switching (WBS) wherein wavelengths are grouped into bands and switched as a single entity can reduce cost and complexity of switching nodes by minimizing the port count. In this paper, we study the effect of wavelength conversion on the performance of WBS networks with reconfigurable multi-granular optical cross-connects (MG-OXC) to satisfy online traffic. Since wavelength conversion is still expensive and can potentially increase the number of used ports in WBS networks, efficient usage of wavelength converters is of practical interest. We propose a novel heuristic algorithm, called waveband assignment with path-graph (WAPG), which takes efficient wavebanding and efficient usage of wavelength converters into consideration when satisfying new lightpath requests. We apply the WAPG algorithm in WBS networks with full, intra-band, or limited number of wavelength converters, and compare with the FirstFit and RandomFit algorithms. Our results indicate that the proposed algorithm performs significantly better in terms of the blocking probability as well as the number of used wavelength converters.

A key problem in optical burst switching (OBS) is to schedule as many bursts as possible on wavelength channels so that the throughput is maximized and the burst loss is minimized. In this paper, we use competitive analysis to analyze the worst-case performance of a large set of scheduling algorithms, called best-effort online scheduling algorithms, for OBS networks, and establish a number of interesting upper and lower bounds on the performance of such algorithms. A surprising discovery is that the worst-case performance of any best-effort online scheduling algorithm is primarily determined by the maximum to minimum burst length ratio, followed by the range of offset time. Furthermore, if all bursts have the same burst length and offset time, all best-effort online scheduling algorithms generate the same optimal solution, regardless how different they may look like. Our analysis can also be extended to some nonbest-effort online scheduling algorithms, such as the well-known Horizon algorithm, and establish similar bounds. Based on the analytic results, we give guidelines for several widely discussed OBS problems, including burst assembly, offset time setting and scheduling algorithm design, and propose a new channel reservation protocol called VFO to improve the worst-case performance. Our simulation shows that it is quite often for an online scheduling algorithm to exhibit its (near) worst-case performance. Thus improving the worst-case performance is essential. Our simulation also suggests that VFO reduces the average burst loss rate by as much as 35%

Optical burst switching (OBS) is a promising paradigm for the next-generation Internet infrastructure. In OBS, a key problem is to schedule bursts on wavelength channels with both fast and bandwidth efficient algorithms so as to reduce burst loss. To date, most scheduling algorithms avoid burst contention locally (or reactively). In this paper, we propose several novel algorithms for scheduling bursts in OBS networks with and without wavelength conversion capability. Our algorithms try to proactively avoid burst contention likely to occur at downstream nodes. The basic idea is to serialize the bursts on an outgoing link to reduce the number of bursts that may arrive at downstream nodes simultaneously (and thus reducing the burst contention and burst loss probability at downstream nodes). This can be accomplished by judiciously delaying locally assembled bursts beyond a pre-determined offset time at an ingress node using the electronic memory. Compared with the existing algorithms, our proposed algorithms can significantly reduce the loss rate while ensuring that maximum delay of a burst does not exceed its prescribed limit.

Optical burst switching (OBS) is a promising paradigm for the next-generation Internet. In OBS, a key problem is to schedule bursts on wavelength channels whose bandwidth may become fragmented with the so-called void (or idle) intervals with both fast and bandwidth efficient algorithms so as to reduce burst loss. To date, only two scheduling algorithms, called Horizon and LAUC-VF, have been proposed, which trade off bandwidth efficiency for fast running time and vice versa, respectively. In this paper, we propose several novel algorithms for scheduling bursts in OBS networks with and without fiber delay lines (FDLs). In networks without FDLs, our proposed Min-SV algorithm can schedule a burst successfully in O(logm) time, where m is the total number of void intervals, as long as there is a suitable void interval. Simulation results suggest that our algorithm achieves a loss rate which is at least as low as the best previously known algorithm LAUC-VF, but can run much faster. In fact, its speed can be almost the same as Horizon (which has a much higher loss rate). In networks with FDLs, our proposed batching FDL algorithm considers a batch of FDLs simultaneously to find a suitable FDL to delay a burst which would otherwise be discarded due to contention, instead of considering the FDLs one by one. The average search time of this algorithm is therefore significantly reduced from that of the existing sequential search algorithms.

Optical burst switching (OBS) is a promising paradigm for the next-generation Internet infrastructure. We study the performance of TCP traffic in OBS networks and in particular, the effect of assembly algorithms on TCP traffic. We describe three assembly algorithms in this paper and compare them using the same TCP traffic input. The results show that the performance of the proposed adaptive-assembly-period (AAP) algorithm is better than that of the min-burstlength-max-assembly-period (MBMAP) algorithm and the fixed-assembly-period (FAP) algorithm in terms of goodput and data loss rate. The results also indicate that burst assembly mechanisms affect the behavior of TCP in that the assembled TCP traffic becomes smoother in the short term, and more suitable for transmission in optical networks.