High Efficiency Video Coding. (HEVC) reflects the new international standardization for digital video coding technology. HEVC achieves higher compression compared to its antecessor at the expense of dramatically increasing coding complexity due to the use of a recursive quadtree to partition every frame to various block sizes, a process called prediction mode. We propose three CNNs based on VGGNet, one CNN for each CU size of 64x64, 32x32, and 16x16, as shown in Figure 1, to predict the quadtree levels for the CU blocks of HEVC reducing its code complexity. The new CNNs simplify the original VGGNet in terms of number of convolutional layers while maintaining the original 3x3 filters. As our model is designed to recognize the quadtree structure of a block of pixels instead of image categories, a shallow version of the VGGNet combined with our CU partition datasets will provide fast and accurate results. The accuracy of the model can be further improved because the input CU size is consistent with the size of CU encoded by HEVC, that avoids losses in the CU texture features. Our CNN models learn from three customized datasets of CU blocks encoded in the specific QP of 32. In this way there is no need to introduce QP as a parameter in the loss function used in other works, and further increase accuracy. Given the success of this idea, in the future models will have separate training for each QP of 22, 27 and 37, respectively.

Pedestrian trajectory prediction is an essential component of autonomous systems and robot navigation. Recent research has shown promising predictive performance by designing prediction networks to model a variety of motion-related features. Different from existing works, our focus is on designing a novel online adaptation framework (OAT-Mem) to exploit the temporal similarities among trajectory samples encountered during testing to improve the prediction accuracy of any such models (i.e., predictors) without knowing the details of these predictors. Our framework consists of two novel modules: an augmented temporal observation-target memory network (ATM) and a certainty-based selector (CS). Inspired by the concept of key-value memory networks [16], ATM is proposed to learn the temporal information from short-term past frames by encoding the trajectory samples of past pedestrians in form of observation-target (i.e., key-value) during testing. In addition, we propose a certainty-based selector (CS) to enhance the predictive ability of our framework under scenarios where there are large temporal dissimilarities between current pedestrians’ movements and those stored in memory. In dynamic scenes, these scenarios commonly occur due to abrupt changes in contexts, such as camera motions, scene contexts, and pedestrians’ behaviors. We extensively evaluate our framework in commonly-used datasets: JAAD [12] and PIE [19] and show that our framework significantly improves the prediction accuracy of state-of-the-art models. Finally, in-depth studies are conducted to show the importance of each proposed component.

We present a new system that can automatically generate input paths (syllabus) for a convolutional neural network to follow through a curriculum learning to improve training performance. Our system utilizes information-theoretic content measures of training samples to form syllabus at training time. We treat every sample as 2D random variable where a data point contained in the sample (such as a pixel) is modelled as an independent and identically distributed random variable (i.i.d) realization. We use several information theory methods to rank and determine when a sample is fed to a network by measuring its pixel composition and its relationship to other samples in the training set. Comparative evaluation of multiple state-of-the-art networks, including, GoogleNet, and VGG, on benchmark datasets demonstrate a syllabus that ranks samples using measures such as Joint Entropy between adjacent samples, can improve learning and significantly reduce the amount of training steps required to achieve desirable training accuracy. We present results that indicate our approach can reduce training loss by as much as a factor of 9 compared to conventional training.

This paper presents a new complexity Reduction method for the diamond search pattern called TZSR based on region labels in HEVC/H.265 video coding. The solution introduces a new image region structure developed from a simplified version of the blob coloring algorithm (image labeling) to HEVC/H.265 video coding. Regions are either whole or part of image objects and normally span several coding tree blocks that are produced during HEVC encoding. Our method executes a complete Diamond Search (DS) for the first block of each region in order to identify the motion vector direction among eight different directions in DS. The motion estimation (ME) for the rest of the blocks in the region will perform a modified DS where only one direction point for various distances will be tested in order to reduce the code complexity. Experimental results demonstrate that the speedup achieved in our solution surpasses the time spent in our blob coloring algorithm. Furthermore, TZSR achieves an average speedup of 42.61% for low delay (LD) configuration and 52.13% for random access (RA) in the encoding time compared to the original ME algorithm in HEVC reference software (HM-16.7) with overall gains in PSNR (Peak Signal-to-Noise Ratio) and bit rate around 18.67% and 0.1 under LD and 28.37% and 0.74 under RA respectively.

Abstract In this paper we offer an integrated flexible system in which two inter‐related models interact. One: the systems environment model that is inherent in all large enterprises and defines the logical enterprise organizations, units, division of work and responsibilities. Two: the Relationship model, which is a dynamic network derived from the data provided by the distributed expertise of the systems environment model representing several types of relationships among the units. This network model can be probed to produce related units with specific relationships to efficiently initiate a change, plan a change, adjudicate and implement a change. Each unit of the network can initiate a proposal for a new change by exploring its effect on the rest of the network; once a proposal for change is explored, its implications are evaluated and approved, a coordinated change is planned and implemented in a timely and organized manner while every constituency has been prepared for the planned change. While the proposed model can easily be adapted for organizations with well‐defined systems environment models, this paper focuses on the design, integration and implementation of the Relationship network and its processes for information technology (IT) within a large telecommunication company with an existing systems environment model. Copyright © 2010 John Wiley & Sons, Ltd.

Abstract The Hydra Parallel Programming System, a new parallel language extension to Java, and its supporting software are described. It is a fairly simple yet powerful language designed to address a number of areas that have not received much attention. One of these areas is the recompilation of parallel programs at runtime to allow a parallel program to adapt to the architecture it is executing on. The first version of this software system focuses on smaller Symmetric Multiprocessing and compatible architectures which are becoming more common. This particular class of machines has a great need for more options in the area of parallel programming among the vastly popular Java language programmers. Hydra programs will run as sequential Java on machines that do not have the parallel support or do not have an implemented Hydra runtime system without requirement of any modifications to the program. This paper describes the language, compares it with other languages (specifically with JOMP, an OpenMP implementation for Java), presents a brief discussion on compiling and executing Hydra programs, presents some sample benchmarks and their performance on three platforms, and concludes with a discussion of issues and future directions for Hydra. Copyright © 2007 John Wiley & Sons, Ltd.

Hydra PPS is a collection of annotations, classes, a runtime, and a compiler designed to provide Java programmers with a fairly simple method of producing programs for Symmetric Multiprocessing (SMP) architectures. This paper introduces the basics of this new system including the basic constructs for this new programming language and the relationship between the Java VM, the compiler, the runtime, and the parallel program. Hydra will exploit parallelism when the underlying architecture supports it and will run as normal sequential Java program when the architecture does not have support for parallelism. Parallelism is expressed through events in Hydra, it is easy to use, and programs run efficiently on parallel architectures.

Explains why the Force parallel programming language has been easily portable between eight different shared memory multiprocessors. The authors show how a two-layer macro processor allows them to hide machine dependencies and to build machine-independent high-level language constructs. The importance of packaging low-level synchronization operations is demonstrated by a proof of mutual exclusion for asynchronous variable operations. The Force constructs enable one to write portable parallel programs largely independent of the number of processes executing them.>

It is shown that in sparse matrices arising from electronic circuits, it is possible to do computations on many diagonal elements simultaneously. A technique for obtaining an ordered compatible set directly from the ordered incompatible table is given. The ordering is based on the Markowitz number of the pivot candidates. This technique generates a set of compatible pivots with the property of generating few fills. A novel heuristic algorithm is presented that combines the idea of an order-compatible set with a limited binary tree search to generate several sets of compatible pivots in linear time. An elimination set for reducing the matrix is generated and selected on the basis of a minimum Markovitz sum number. The parallel pivoting technique presented is a stepwise algorithm and can be applied to any submatrix of the original matrix. Thus, it is not a preordering of the sparse matrix and is applied dynamically as the decomposition proceeds. Parameters are suggested to obtain a balance between parallelism and fill-ins. Results of applying the proposed algorithms on several large application matrices using the HEP multiprocessor are presented and analyzed.>