The IETF 6TiSCH Working Group proposed a highly reliable and low-power industrial wireless network protocol stack, which integrates IEEE 802.15.4e Time Slotted Channel Hopping(TSCH) with IPv6 protocols. Scheduling is crucial in the 6TiSCH protocol stack, as it determines when a node sends or receives network packets on specific timeslots and channels. However, the current resource scheduling suffers from random timeslot selection, channel collisions, and high energy consumption, all of which degrade network performance. In this article, we propose an innovative scheduling function consisting of three core components. First, a multi-weight quality evaluation is proposed to enable precise timeslot selection, improving resource allocation efficiency. Second, an improved Q-Learning-based scheduling mechanism is proposed to generate optimal parameters for multi-weight evaluations, which is based on the global network state implemented at the edge router. Third, a pseudo-random channel selection strategy is proposed to effectively reduce channel congestion and interference. Extensive experimental results demonstrate that our function achieves a latency improvement of up to 55% compared to state-of-the-art low-latency scheduling, and along with an enhancement of up to 6% in the packet delivery ratio.

The growing memory demands of modern applications have driven the adoption of far memory technologies in data centers to provide cost-effective, high-capacity memory solutions. However, far memory presents new performance challenges because its access latencies are significantly longer and more variable than local DRAM. For applications to achieve acceptable performance on far memory, a high degree of memory-level parallelism (MLP) is needed to tolerate the long access latency. While modern out-of-order processors are capable of exploiting a certain degree of MLP, they are constrained by resource limitations and hardware complexity. The key obstacle is the synchronous memory access semantics of traditional load/store instructions, which occupy critical hardware resources for a long time. The longer far memory latencies exacerbate this limitation. This article proposes a set of Asynchronous Memory Access Instructions (AMI) and its supporting function unit, Asynchronous Memory Access Unit (AMU), inside contemporary Out-of-Order Core. AMI separates memory request issuing from response handling to reduce resource occupation. Additionally, AMU architecture supports up to several hundreds of asynchronous memory requests through re-purposing a portion of L2 Cache as scratchpad memory (SPM) to provide sufficient temporal storage. Together with a coroutine-based programming framework, this scheme can achieve significantly higher MLP for hiding far memory latencies. Evaluation with a cycle-accurate simulation shows AMI achieves 2.42× speedup on average for memory-bound benchmarks with 1μs additional far memory latency. Over 130 outstanding requests are supported with 26.86× speedup for GUPS (random access) with 5 μs latency. These demonstrate how the techniques tackle far memory performance impacts through explicit MLP expression and latency adaptation.

Memory-intensive applications, such as in-memory databases, caching systems, and key-value stores, are increasingly demanding larger main memory to fit their working sets. Conventional swapping can enlarge the memory capacity by paging out inactive pages to backend stores. However, existing swapping solutions suffer several performance and compatibility issues, making them unsuitable for high-concurrency and memory-intensive applications. In this article, we redesign the swapping system and propose Lightswap, a high-performance user-space swapping solution that supports paging with both local SSDs and remote memories. First, to avoid kernel involvement, we propose to leverage the extended Berkeley packet filter (eBPF) for handling page faults (PFs) in user space and further eliminate the heavy I/O stack with the help of user-space I/O drivers. Then, we co-design the PF handling with lightweight thread (LWT) scheduling to improve system throughput and reduce the end-to-end PF latency. Finally, we propose a try-catch framework in Lightswap to deal with swap-in errors which have been exacerbated by the scaling in process technology. We implement Lightswap in our production-level system and evaluate it with various benchmarks. Results show that Lightswap achieves scalable PF notification latency ($4 \mu \text{s}$under 128 LWTs), reduces the PF handling latency by 3–5 times, and improves the throughput of memcached by more than 40% compared with the state-of-art swapping systems.

The traditional center/surround Retinex enhancement algorithms only consider the illuminance. So it may amplify the noise. However image details will be lost in the subsequent denoising process. In this paper, we propose an illuminance-reflectance model(IRMNE) which can denoise the image taken in the low light and enhance it at the same time. IRMNE can overcome the limitations of traditional algorithms. The experiment results indicate the validity and advantage of our algorithm. Compared with other algorithms, IRMNE makes a better balance between image enhancement and image denoising, which helps to get more reliable color restoration results.

Despite very significant efforts on smartphonebased indoor localization, highly accurate and practical method remains an open problem. To guarantee accuracy, robustness, and practicality, in this paper, we propose SITE, a novel scheme uses acoustic Signal and phone Images to achieve accurate and robust indoor posiTion systEm. Our key observation is that if the simultaneously computed locations according to different sets of acoustic sources vary small, the positioning result is close to the true physical location. Based on the pre-deployed acoustic sources, SITE first tracks the direction of smartphone relative to an individual acoustic source according to proactively generated doppler effect in rough horizontal plane. Given m (m >= 3) acoustic sources, SITE can compute the relative coordinate of the phone in floor plan. By respectively selecting different sets of acoustic sources to compute the related coordinates, SITE can make sure whether these positioning results satisfy the requirement of positioning accuracy. If not, using images captured by phone camera, SITE exploits the synergy between its acoustic-based localization (coarse-grained) and the relative positions in reconstructed 3D point cloud by VisualSFM technique to refine the positioning result using acoustic signals. We have built a prototype of the SITE system and conducted evaluations in real testbed. Experimental results show that SITE is excellent in accuracy, robust and valuable in practical application.

Nighttime traffic surveillance is difficult due to insufficient and unstable appearance information and strong background interference. We present in this paper a robust nighttime vehicle detection system by detecting, tracking, and grouping headlights. First, we train AdaBoost classifiers for headlights detection to reduce false alarms caused by reflections. Second, to take full advantage of the complementary nature of grouping and tracking, we alternately optimize grouping and tracking. For grouping, motion features produced by tracking are used by headlights pairing. We use a maximal independent set framework for effective pairing, which is more robust than traditional pairing-by-rules methods. For tracking, context information provided by pairing is employed by multiple object tracking. The experiments on challenging datasets and quantitative evaluation show promising performance of our method.

Interactive segmentation, in which a user provides a bounding box to an object of interest for image segmentation, has been applied to a variety of applications in image editing, crowdsourcing, computer vision, and medical imaging. The challenge of this semi-automatic image segmentation task lies in dealing with the uncertainty of the foreground object within a bounding box. Here, we formulate the interactive segmentation problem as a multiple instance learning (MIL) task by generating positive bags from pixels of sweeping lines within a bounding box. We name this approach MILCut. We provide a justification to our formulation and develop an algorithm with significant performance and efficiency gain over existing state-of-the-art systems. Extensive experiments demonstrate the evident advantage of our approach.

The high demand from railway safety has promoted the sensing technology to avoid manual inspection that is error-prone, low efficiency, labor-intensive and high costs. However, few works have emphasized railway environmental monitoring so far which is equally important in railway safety maintenance. In this paper, we propose a railway scene rendering system for virtual maneuver and inspection. Based on the route panoramas extracted from a real train-borne video, we generate the compact and high fidelity scenes over a long distance route in cyberspace. We analyze the relation between train motion and the video sampling in order to obtain the route panoramas of the railways at the just-sampling rate. Then, we render scene tunnels based on the route panorama in a virtual space for interactive navigation, inspection, and scene indexing. The experimental results show remarkable advantages in the data storage and automatic inspection. The compact data format with sufficient visual information can be transmitted and released in cyberspace for fast access, maneuver, and automatic retrieval in railway environment monitoring.

With the rapid development of high-speed railway, the safety running of trains becomes extremely urgent. Video is a direct and effective manner for railway environment monitoring. In this paper, we propose a route panoramic representation to produce a virtual environment of the railway scene from a train-borne video, which provides an all-round display for the entire length of railway. We generate the route panorama from the forward motion video and then render 3D scene. The experimental results show many advantages of the representation in the large data storage, browsing, and examination. It will be used for railway safety checking, railway facility inspection and virtual sightseeing from the train in the future.

Compressing a query image's signature via vocabulary coding is an effective approach to low bit rate mobile visual search. State-of-the-art methods concentrate on offline learning a codebook from an initial large vocabulary. Over a large heterogeneous reference database, learning a single codebook may not suffice for maximally removing redundant codewords for vocabulary based compact descriptor. In this paper, we propose to learn multiple codebooks (m-Codebooks) for extremely compressing image signatures. A query-specific codebook (q-Codebook) is online generated at both client and server sides by adaptively weighting the off-line learned multiple codebooks. The q-Codebook is subsequently employed to quantize the query image for producing compact, discriminative, and scalable descriptors. As q-Codebook may be simultaneously generated at both sides, without transmitting the entire vocabulary, only small overhead (e.g. codebook ID and codeword 0/1 index) is incurred to reconstruct the query signature at the server end. To fulfill m-Codebooks and q-Codebook, we adopt a Bi-layer Sparse Coding method to learn the sparse relationships of codewords vs. codebooks as well as codebooks vs. query images via l1 regularization. Experiments on benchmarking datasets have demonstrated the extremely small descriptor's supervior performance in image retrieval.

User-generated tags associated with images from social media (e.g., Flickr) provide valuable textual resources for image classification. However, the noisy and huge tag vocabulary heavily degrades the effectiveness and efficiency of state-of-the-art image classification methods that exploited auxiliary web data. To alleviate the problem, we introduce a Sparse Tag Patterns (STP) model to discover sparsity constrained co-occurrence tag patterns from large scale user contributed tags among social data. To fulfill the compactness and discriminability, we formulate STP as a problem of minimizing a quadratic loss function regularized by the bi-layer l1norm. We treat the learned STP as alternative intermediate semantic image feature and verify its superiority within a search-based image classification framework. Experiments on 240K social images associated with millions of tags have demonstrated encouraging performance of the proposed method compared to the state-of-the-art.

User-given tags associated with social images from photosharing websites (e.g., Flickr) are valuable auxiliary resources for the image tagging task. However, social images often suffer from noisy and incomplete tags, heavily degrading the effectiveness of previous image tagging approaches. To alleviate the problem, we introduce a Sparse Tag Patterns (STP) model to discover noiseless and complementary cooccurrence tag patterns from large scale user contributed tags among auxiliary web data. To fulfill the compactness and discriminability, we formulate the STP model as a problem of minimizing quadratic loss function regularized by bi-layer ℓ1norm. We treat the learned STP as a universal knowledge base and verify its superiority within a data-driven image tagging framework. Experimental results over 1 million auxiliary data demonstrate superior performance of the proposed method compared to the state-of-the-art.

Focusing on the fact that the grid users may require various QoS and have quite different preferences on each QoS, a User Preference Driven Approach(UPDA) is proposed to schedule grid tasks with multi-QoS constraints. In UPDA, user's needs and preferences are modeled into matrixes to serve as the basis of scheduling and Analytical Hierarchy Process(AHP) method is adopted to perform a pair wise comparison between different kinds of QoS demands to gain the preference on each QoS. Furthermore, UPDA tries to offer optimal performance on the multi-QoS targets with the consideration of the resource utilization, which is just in compliance with the ground rule of fully utilizing of spare resource in grid. Simulation results confirm that UPDA can not only better satisfy the diverse requirements of users, but also perform well in term of the resource utilization rate from the system's perspectives.

Scheduling mechanism in cloud computing environment concerns more about the needs of users due to the commercial purpose of cloud computing. In response to this user-oriented feature, an Adaptive Resource Scheduling Mechanism based on User Behavior Feedback(ARSM-UBF) is developed to schedule cloud resources from the perspective of users. In ARSM-UBF, information collected from scheduling interaction is adopted as the reflection of user behavior and by introducing the relevance feedback mechanism, user behavior information is integrated into the resource scheduling and continuous feedback is proceeded on the following scheduling process to make scheduling results closer to the user's subjective perception. Furthermore, ARSM-UBF gives attention to the efficiency of cloud system. Simulation results confirm that this user behavior feedback based scheduling mechanism is much more efficient since it can not only provide guaranteed services for users with diverse requirements, but also improve the resource utilization rate from the system's perspectives.

Inspired by information processing of complex cells in visual cortex, we present a simple system for fast and robust feature extraction. Our method includes an unsupervised algorithm for learning invariant descriptors from data, and an architecture for the task of digit recognition. The proposed algorithm is not only efficient that training can be accomplished in a few iterations, but can map test data into invariant representations directly, in contrast to most existing generative model, which must perform inference by minimizing energy functions. The simulation results on the well known MNIST database show that these learnt descriptors demonstrate a clear topography with similar properties of complex cells, and extract features that are invariant to minor variations of input data. Recognition experiments also show that the learnt invariant feature descriptors improve the accuracy than classical feature descriptors and yield comparable classification results.

Experts-Shift is a novel statistical framework for keyframe-based video segmentation. Compared to existing video segmentation techniques with simple color models, our method proposes a probability mixture model coupling strong image classifiers (experts) with latent spatial configuration. In order to propagate image labels to the successive frames, our algorithm track all experts jointly by a efficient MCMC sampler with their relations modeled by MRFs. This algorithm is capable to handle overlapping color distribution, ambiguous image boundaries, large displacement in challenging scenario with a solid foundation of both generative modeling and discriminative learning. Experiment shows our algorithm achieves high quality results and need less supervision than previous work.

Image blurring with Gaussian kernel can also be modeled as an heat diffusion partial differential equation (PDE), therefore, image deblur is an inverse problem of PDE. In this paper, we proposed a novel image deblur method: regularized backward heat diffusion (RBHD) PDE, based on inverse problem and PDE theories. We have derived the concrete regularization forms: low-pass filtering or adding high order derivative terms, provided several kinds of RBHD PDE, and analyzed the relationship between the optimal cut-off frequency and the estimated forward diffusion time. Compared to the traditional energy based methods, e.g. Wiener filter, our method is more flexible and extensible, and the experiments shows that RBHD PDE achieves much better results both with and without the exact kernel width, and is more suitable as an experimental blind deblurring method for Gaussian blurry image.

In computer vision, the bag-of-visual words(BOV) approach has been shown to yield state-of-the-art results. To improve BOV model, we use sparse codes of SIFT features instead of previous vector quantization (VQ) such as k-means, due to more quantization errors of VQ. And as local features in most categories have spatial dependence in real world, we use neighbor features of one local feature as its implicit local spatial relationship (ILSR). This paper proposes an object categorization algorithm which integrate implicit local spatial relationship with its appearance features based on sparse codes of SIFT to form two sources of information for categorization. The algorithm is applied in Caltech-101 and Caltech-256 datasets to validate its effectiveness. The experimental results show its good performance.

This paper presents an interactive image segmentation framework which is ultra-fast and accurate. Our framework, termed "CO3", consists of three components: COupled representation, COnditional model and COnvex inference. (i) In representation, we pose the segmentation problem as partitioning an image domain into regions (foreground vs. background) or boundaries (on vs. off) which are dual but simultaneously compete with each other. Then, we formulate segmentation process as a combinatorial posterior ratio test in both the region and boundary partition space. (ii) In modeling, we use discriminative learning methods to train conditional models for both region and boundary based on interactive scribbles. We exploit rich image features at multi-scales, and simultaneously incorporate user's intention behind the interactive scribbles. (iii) In computing, we relax the energy function into an equivalent continuous form which is convex. Then, we adopt the Bregman iteration method to enforce the "coupling" of region and boundary terms with fast global convergence. In addition, a multigrid technique is further introduced, which is a coarse-to-fine mechanism and guarantees both feature discriminativeness and boundary preciseness by adjusting the size of image features gradually.

Slow Feature Analysis (SFA) is an unsupervised algorithm by extracting the slowly varying features from time series and has been used to pattern recognition successfully. Based on SFA, this paper develops a new algorithm, Slow Feature Discriminant Analysis (SFDA), which can maximize the temporal variation of between-class time series, and minimize the temporal variation of within-class time series simultaneously. Due to adoption of discrimination power, the performance on pattern recognition is improved compared to SFA. The experiments results on MNIST digit handwritten database also show that the proposed algorithm is in particular attractive.

The paper presents a dimensionality reduction method called Neighboring Discriminant Analysis (NDA) and its kernel extension to improve face recognition performance. We take into account both the data distribution and class label information. We describe the connection of two data as neighboring or non-neighboring, together with whether the pair are from the same class or belong to different classes by utilizing the Graph Embedding framework as a tool. The compactness graph is constructed by connecting each data vertex with its neighboring data of the same class, while the penalty graph connects the rest data pairs, i.e. the data pairs which are not from the same class or are non-neighboring. NDA algorithm can map the original high dimensional space to a reduced low dimensional space, which compact the neighboring data from the same class and simultaneously separate the data far away from each other or belong to different classes. Real face recognition experiment shows NDA and its kernel extension outperforms LDA etc.

Object detection in clutter or occlusion is a hard problem in computer vision. We propose an object detection method based on contour grouping. Two stages are included: a novel distance transform is applied to match templates to the test image so that candidates and locations of the object are obtained; verification using shape manifold is performed to preclude outliers and identify the prior. We use the prior combined with bottom-up edge information to produce the final grouping result. Our contribution lies in two aspects: one is the novel distance transform saves much searching space; the other is introducing shape manifold in verifying candidates of grouping. Experiments show our method achieves considerable accuracy in occlusion and background clutter. Specially, the only feature used is edge and contour rather than combination of multi features.

Semi-supervised learning has received much attention recently. Co-training is a kind of semi-supervised learning method which uses unlabeled data to improve the performance of standard supervised learning algorithms. A novel co-training style algorithm, RASCO (for RAndom Subspace CO-training), is proposed which uses stochastic discrimination theory to extend co-training to multi-view situation. The accuracy and generalizability of RASCO are analyzed. The influences of the parameters of RASCO are discussed. Experiments on UCI data set demonstrate that RASCO is more effective than other co-training style algorithms.

In order to improve the performance of content-based image retrieval (CBIR) systems, the 'semantic gap' between the low-level visual features and the high-level semantic features attracts more and more research interest. We propose an approach to describe and to extract the fuzzy color semantics. According to human color perception model, we utilize the linguistic variable to describe the image color semantics, so it becomes possible to depict the image in linguistic expression such as mostly red. Furthermore, we apply the feedforward neural network to model the vagueness of human color perception and to extract the fuzzy semantic feature vector. Our experiments show that the color semantic features have good accordance with the human perception, and also have good retrieval performance. In some extent, our approach shows the potential to reduce the semantic gap in CBIR.

Selective attention is an important mechanism in human visual system. Existing attention models concentrate on giving a series of sight transferring. In this paper, we extend selective attention to guide contour extraction prepared for perceptual grouping. Selective attention functions in two aspects. One is to compute each pixel's salient value in an image. The other is to extract salient contours. Compared with other contour extraction algorithms, our model gets better result when the same edge detect cue is used. Experiments and quantitative analysis testify our model's good performance contour extraction and contour grouping.

One of the big challenges faced by content-based image retrieval (CBIR) is the 'semantic gap' between the visual features and the richness of human semantics for image content. We put forward a neural network approach to extract the image fuzzy semantics ground on linguistic expression based image description framework (LEBID). We utilize the linguistic variable to depict the texture semantics according to Tamura texture model, so we can describe the image in linguistic expression such as coarse, very line-like. Moreover, we use feedforward neural network (NN) to model the vagueness of human visual perception and to extract the fuzzy semantic feature. Our experiments demonstrate that NN outperforms other method such as genetic algorithm on the complexity of model, and it also achieves good retrieval performance.

This letter develops a simple and effective whitening method for flattening the power-law power spectrum of face images and combines the whitening technique and PCA/ICA for face recognition. The whitening parameter in the whitening filter is adaptive to the input data so as to make the power spectrum of the whitened results as flat as possible. This proposed method is computationally saving for using fast Fourier transform algorithms. The transformed ldquowhitenedfacesrdquo are applied to face recognition with PCA and ICA coding strategies, respectively. The results show that the whitening processing will bring better recognition performances for both PCA and ICA representations.

In conventional feature extraction based on independent component analysis (ICA), feature selection and dimensional reduction are carried out only through PCA preprocessing, so the importance of independent components is not taken into consideration. In order to overcome this problem, a new ICA feature selection based on genetic algorithm is proposed in this paper. To demonstrate its effectiveness, recognition experiments is performed for face recognition and iris recognition.

In this paper a neural network for solving partial differential equations (PDE) is described. The activation functions of the hidden nodes are the radial basis functions (RBF) whose parameters are learnt by a two-stage gradient descent strategy. A new growing radial basis functions-node insertion strategy with different radial basis functions is used in order to improve the net performances. The learning strategy is able to save computational time and memory space because of the selective growing of nodes whose activation functions consist of different radial basis functions. An analysis of the learning capabilities and a comparison of the net performances with other approaches have been performed. It is shown that the resulting network improves the approximation results.