Object contour detection is a task of extracting the shape created by the boundaries between objects in an image. Conventional methods limit the detection targets to specific categories, or miss-detect edges of patterns inside an object. We propose a new method to represent a contour image where the pixel value is the distance to the boundary. Contour detection becomes a regression problem that estimates this contour image. A deep convolutional network for contour estimation is combined with stereo vision to detect unspecified object contours. Furthermore, thanks to similar inference targets and common network structure, we propose a network that simultaneously estimates both contour and disparity with fully shared weights. As a result of experiments, the multi-tasking network drew a good precision-recall curve, and F-measure was about 0.833 for FlyingThings3D dataset. L1 loss of disparity estimation for the dataset was 2.571. This network reduces the amount of calculation and memory capacity by half, and accuracy drop compared to the dedicated networks is slight. Then we quantize both weights and activations of the network to 3-bit. We devise a dedicated hardware architecture for the quantized CNN and implement it on an FPGA. This circuit uses only internal memory to perform forward propagation calculations, that eliminates high-power external memory accesses. This circuit is a stall-free pixel-by-pixel pipeline, and performs 8 rows, 16 input channels, 16 output channels, 3 by 3 pixels convolution calculations in parallel. The convolution calculation performance at the operating frequency of 250 MHz is 9 TOPs/s.

We propose an attention mechanism in deep learning networks for gender recognition using the gaze distribution of human observers when they judge the gender of people in pedestrian images. Prevalent attention mechanisms spatially compute the correlation among values of all cells in an input feature map to calculate attention weights. If a large bias in the background of pedestrian images (e.g., test samples and training samples containing different backgrounds) is present, the attention weights learned using the prevalent attention mechanisms are affected by the bias, which in turn reduces the accuracy of gender recognition. To avoid this problem, we incorporate an attention mechanism called gaze-guided self-attention (GSA) that is inspired by human visual attention. Our method assigns spatially suitable attention weights to each input feature map using the gaze distribution of human observers. In particular, GSA yields promising results even when using training samples with the background bias. The results of experiments on publicly available datasets confirm that our GSA, using the gaze distribution, is more accurate in gender recognition than currently available attention-based methods in the case of background bias between training and test samples.

Exponential growth in data volumes has promoted widespread interest in data-selective adaptive algorithms. In a pioneering work, Diniz developed the data-selective least mean square (DS-LMS) algorithm, which is able to reduce specific quantities of computation data without compromising performance. Note however that the existing framework fails to consider the issue of impulse noise (IN), which can greatly undermine the benefits of reduced computation. In this letter, we present an error-based IN detection algorithm for implementation in conjunction with the DS-LMS algorithm. Numerical evaluations confirm the effectiveness of our proposed IN-tolerant DS-LMS algorithm.

Frequency-hopping sequence (FHS) sets with low-hit-zone (LZH) can be well applied in quasi-synchronous (QS) frequency-hopping multiple-access (FHMA) systems to reduce the mutual interference among different users. On the other hand, LHZ-FHS sets with wide-gap (WG) property can effectively resist the broadband blocking interference, the single frequency narrowband interference, the multipath fading and the tracking interference. In this letter, a new family of WG-LHZ-FHS sets is constructed. Besides, these new WG-LHZ-FHS sets possess optimal average periodic Hamming correlation (APHC) properties.

In this paper, we propose a novel Hybrid-Hierarchical spatial-aerial-Terrestrial Edge-Centric (H2TEC) for the space-air integrated Internet of Things (IoT) networks. (H2TEC) comprises unmanned aerial vehicles (UAVs) that act as mobile fog nodes to provide the required services for terminal nodes (TNs) in cooperation with the satellites. TNs in (H2TEC) offload their generated tasks to the UAVs for further processing. Due to the limited energy budget of TNs, a novel task allocation protocol, named TOP, is proposed to minimize the energy consumption of TNs while guaranteeing the outage probability and network reliability for which the transmission rate of TNs is optimized. TOP also takes advantage of the energy harvesting by which the low earth orbit satellites transfer energy to the UAVs when the remaining energy of the UAVs is below a predefined threshold. To this end, the harvested power of the UAVs is optimized alongside the corresponding harvesting time so that the UAVs can improve the network throughput via processing more bits. Numerical results reveal that TOP outperforms the baseline method in critical situations that more power is required to process the task. It is also found that even in such situations, the energy harvesting mechanism provided in the TOP yields a more efficient network throughput.

In this work, we consider a remote estimation system where a remote controller estimates the status of heterogeneous sensing devices with the information delivered over wireless channels. Status of heterogeneous devices changes at different speeds. With limited wireless resources, estimating as accurately as possible requires careful design of status update schemes. Status update schemes can be divided into two classes: centralized and decentralized. In centralized schemes, a central scheduler coordinates devices to avoid potential collisions. However, in decentralized schemes where each device updates on its own, update decisions can be made by using the current status which is unavailable in centralized schemes. The relation between these two schemes under the heterogeneous devices case is unclear, and thus we study these two schemes in terms of the mean square error (MSE) of the estimation. For centralized schemes, since the scheduler does not have the current status of each device, we study policies where the scheduling decisions are based on age of information (AoI), which measures the staleness of the status information held in the controller. The optimal scheduling policy is provided, along with the corresponding MSE. For decentralized schemes, we consider deviation-based policies with which only devices with estimation deviations larger than prescribed thresholds may update, and the others stay idle. We derive an approximation of the minimum MSE under the deviation-based policies and show that it is e/3 of the minimum MSE under the AoI-based policies. Simulation results further show that the actual minimum MSEs of these two policies are even closer than that shown by the approximation, which indicates that the cost of collision in the deviation-based policy cancels out the gain from exploiting status deviations.

We proposed a population-based metaheuristic called the spy algorithm for solving optimization problems and evaluated its performance. The design of our spy algorithm ensures the benefit of exploration and exploitation as well as cooperative and non-cooperative searches in each iteration. We compared the spy algorithm with genetic algorithm, improved harmony search, and particle swarm optimization on a set of non-convex functions that focus on accuracy, the ability of detecting many global optimum points, and computation time. From statistical analysis results, the spy algorithm outperformed the other algorithms. The spy algorithm had the best accuracy and detected more global optimum points within less computation time, indicating that our spy algorithm is more robust and faster then these other algorithms.

Front video and sensor data captured by vehicle-mounted event recorders are used for not only traffic accident evidence but also safe-driving education as near-miss traffic incident data. However, most event recorder (ER) data shows only regular driving events. To utilize near-miss data for safe-driving education, we need to be able to easily and rapidly locate the appropriate data from large amounts of ER data through labels attached to the scenes/events of interest. This paper proposes a method that can automatically identify near-misses with objects such as pedestrians and bicycles by processing the ER data. The proposed method extracts two deep feature representations that consider car status and the environment surrounding the car. The first feature representation is generated by considering the temporal transitions of car status. The second one can extract the positional relationship between the car and surrounding objects by processing object detection results. Experiments on actual ER data demonstrate that the proposed method can accurately identify and tag near-miss events.

Traffic congestion in road networks has been studied as the congestion game in game theory. In the existing work, the road usage by each agent was assumed to be static during the whole time horizon of the agent's travel, as in the classical congestion game. This assumption, however, should be reconsidered because each agent sequentially uses roads composing the route. In this paper, we propose a multi-agent distributed route selection scheme based on a gradient descent method considering the time-dependency among agents' road usage for vehicular networks. The proposed scheme first estimates the time-dependent flow on each road by considering the agents' probabilistic occupation under the first-in-first-out (FIFO) policy. Then, it calculates the optimal route choice probability of each route candidate using the gradient descent method and the estimated time-dependent flow. Each agent finally selects one route according to the optimal route choice probabilities. We first prove that the proposed scheme can exponentially converge to the steady-state at the convergence rate inversely proportional to the product of the number of agents and that of individual route candidates. Through simulations under a grid-like network and a real road network, we show that the proposed scheme can improve the actual travel time by 5.1% and 2.5% compared with the conventional static-flow based approach, respectively. In addition, we demonstrate that the proposed scheme is robust against incomplete information sharing among agents, which would be caused by its low penetration ratio or limited transmission range of wireless communications.

We present detailed measurements and analysis of the guided acoustic wave Brillouin scattering (GAWBS)-induced depolarization noise in a multi-core fiber (MCF) used for a digital coherent optical transmission. We first describe the GAWBS-induced depolarization noise in an uncoupled four-core fiber (4CF) with a 125μm cladding and compare the depolarization noise spectrum with that of a standard single-mode fiber (SSMF). We found that off-center cores in the 4CF are dominantly affected by higher-order TRn,m modes rather than the TR2,m mode unlike in the center core, and the total power of the depolarization noise in the 4CF was almost the same as that in the SSMF. We also report measurement results for the GAWBS-induced depolarization noise in an uncoupled 19-core fiber with a 240μm cladding. The results indicate that the amounts of depolarization noise generated in the cores are almost identical. Finally, we evaluate the influence of GAWBS-induced polarization crosstalk (XT) on a coherent QAM transmission. We found that the XT limits the achievable multiplicity of the QAM signal to 64 in a transoceanic transmission with an MCF.

In Bitcoin and Ethereum, nodes require a large storage capacity to maintain all of the blockchain data such as transactions. As of September 2021, the storage size of the Bitcoin blockchain has expanded to 355 GB, and it has increased by approximately 50 GB every year over the last five years. This storage requirement is a major hurdle to becoming a block proposer or validator. We propose an architecture called Trail that allows nodes to hold all blocks in a small storage and to generate and validate blocks and transactions. A node in Trail holds all blocks without transactions, UTXOs or account balances. The block size is approximately 8 kB, which is 100 times smaller than that of Bitcoin. On the other hand, a client who issues transactions needs to hold proof of its assets. Thus, compared to traditional blockchains, clients must store additional data. We show that proper data archiving can keep the account device storage size small. Then, we propose a method of executing smart contracts in Trail using a threshold signature. Trail allows more users to be block proposers and validators and improves the decentralization and security of the blockchain.

Unsupervised domain adaptation (DA) is a challenging machine learning problem since the labeled training (source) and unlabeled testing (target) sets belong to different domains and then have different feature distributions, which has recently attracted wide attention in micro-expression recognition (MER). Although some well-performing unsupervised DA methods have been proposed, these methods cannot well solve the problem of unsupervised DA in MER, a. k. a., cross-domain MER. To deal with such a challenging problem, in this letter we propose a novel unsupervised DA method called Joint Patch weighting and Moment Matching (JPMM). JPMM bridges the source and target micro-expression feature sets by minimizing their probability distribution divergence with a multi-order moment matching operation. Meanwhile, it takes advantage of the contributive facial patches by the weight learning such that a domain-invariant feature representation involving micro-expression distinguishable information can be learned. Finally, we carry out extensive experiments to evaluate the proposed JPMM method is superior to recent state-of-the-art unsupervised DA methods in dealing with cross-domain MER.

Zero-shot learning (ZSL) aims to classify images of unseen classes by learning relationship between visual and semantic features. Existing works have been improving recognition accuracy from various approaches, but they employ computationally intensive algorithms that require iterative optimization. In this work, we revisit the primary approach of the pattern recognition, ı.e., nearest neighbor classifiers, to solve the ZSL task by an extremely simple and fast way, called SimpleZSL. Our algorithm consists of the following three simple techniques: (1) just averaging feature vectors to obtain visual prototypes of seen classes, (2) calculating a pseudo-inverse matrix via singular value decomposition to generate visual features of unseen classes, and (3) inferring unseen classes by a nearest neighbor classifier in which cosine similarity is used to measure distance between feature vectors. Through the experiments on common datasets, the proposed method achieves good recognition accuracy with drastically small computational costs. The execution time of the proposed method on a single CPU is more than 100 times faster than those of the GPU implementations of the existing methods with comparable accuracies.

There exists a great demand for automatic anomaly detection in industrial world. The anomaly has been defined as a group of samples that rarely or never appears. Given a type of products, one has to collect numerous samples and train an anomaly detector. When one diverts a model trained with old types of products with sufficient inventory to the new type, one can detect anomalies of the new type before a production line is established. However, because of the definition of the anomaly, a typical anomaly detector considers the new type of products anomalous even if it is consistent with the standard. Given the above practical demand, this study propose a novel problem setting, few-shot anomaly detection, where an anomaly detector trained in source domains is adapted to a small set of target samples without full retraining. Then, we tackle this problem using a hierarchical probabilistic model based on deep learning. Our empirical results on toy and real-world datasets demonstrate that the proposed model detects anomalies in a small set of target samples successfully.

The Proof of Stake (PoS) protocol is one of the consensus algorithms for blockchain, in which the integrity of a new block is validated according to voting by nodes called validators. However, due to validator-oriented voting, voting results are likely to be false when the number of validators with wrong votes increases. In the PoS protocol, validators are motivated to vote correctly by reward and penalty mechanisms. With such mechanisms, validators who contribute to correct consensuses are rewarded, while those who vote incorrectly are penalized. In this paper, we consider an incentivization mechanism based on the voting profile of a validator, which is estimated from the voting history of the validator. In this mechanism, the stake collected due to the penalties are redistributed to validators who vote correctly, improving the incentive of validators to contribute to the system. We evaluate the performance of the proposed mechanism by computer simulations, investigating the impacts of system parameters on the estimation accuracy of the validator profile and the amount of validator's stake. Numerical results show that the proposed mechanism can estimate the voting profile of a validator accurately even when the voting profile dynamically changes. It is also shown that the proposed mechanism gives more reward to validators who vote correctly with high voting profile.

The recommend system has been widely used in many web application areas such as e-commerce services. With the development of the recommend system, the HIN modeling method replaces the traditional bipartite graph modeling method to represent the recommend system. But several studies have already showed that recommend system is vulnerable to shilling attack (injecting attack). However, the effectiveness of how traditional shilling attack has rarely been studied directly in the HIN model. Moreover, no study has focused on how to enhance shilling attacks against HIN recommend system by using the high-level semantic information. This work analyzes the relationship between the high-level semantic information and the attacking effects in HIN recommend system. This work proves that attack results are proportional to the high-level semantic information. Therefore, we propose a heuristic attack method based on high-level semantic information, named Semantic Shilling Attack (SSA) on a HIN recommend system (HERec). This method injects a specific score into each selected item related to the target in semantics. It ensures transmitting the misleading information towards target items and normal users, and attempts to interfere with the effect of the recommend system. The experiment is dependent on two real-world datasets, and proves that the attacking effect is positively correlate with the number of meta-paths. The result shows that our method is more effective when compared with existing baseline algorithms.

Cross-project defect prediction (CPDP) is a hot research topic in recent years. The inconsistent data distribution between source and target projects and lack of labels for most of target instances bring a challenge for defect prediction. Researchers have developed several CPDP methods. However, the prediction performance still needs to be improved. In this paper, we propose a novel approach called Joint Domain Adaption and Pseudo-Labeling (JDAPL). The network architecture consists of a feature mapping sub-network to map source and target instances into a common subspace, followed by a classification sub-network and an auxiliary classification sub-network. The classification sub-network makes use of the label information of labeled instances to generate pseudo-labels. The auxiliary classification sub-network learns to reduce the distribution difference and improve the accuracy of pseudo-labels for unlabeled instances through loss maximization. Network training is guided by the adversarial scheme. Extensive experiments are conducted on 10 projects of the AEEEM and NASA datasets, and the results indicate that our approach achieves better performance compared with the baselines.

We have utilized single-pixel imaging and deep-learning to solve the privacy-preserving problem in gesture recognition for interactive display. Silhouette images of hand gestures were acquired by use of a display panel as an illumination. Reconstructions of gesture images have been performed by numerical experiments on single-pixel imaging by changing the number of illumination mask patterns. For the training and the image restoration with deep learning, we prepared reconstructed data with 250 and 500 illuminations as datasets. For each of the 250 and 500 illuminations, we prepared 9000 datasets in which original images and reconstructed data were paired. Of these data, 8500 data were used for training a neural network (6800 data for training and 1700 data for validation), and 500 data were used to evaluate the accuracy of image restoration. Our neural network, based on U-net, was able to restore images close to the original images even from reconstructed data with greatly reduced number of illuminations, which is 1/40 of the single-pixel imaging without deep learning. Compared restoration accuracy between cases using shadowgraph (black on white background) and negative-positive reversed images (white on black background) as silhouette image, the accuracy of the restored image was lower for negative-positive-reversed images when the number of illuminations was small. Moreover, we found that the restoration accuracy decreased in the order of rock, scissor, and paper. Shadowgraph is suitable for gesture silhouette, and it is necessary to prepare training data and construct neural networks, to avoid the restoration accuracy between gestures when further reducing the number of illuminations.

Sharding is a solution to the blockchain scalability problem. A sharded blockchain divides consensus nodes (validators) into groups called shards and processes transactions separately to improve throughput and latency. In this paper, we analyze the rational behavior of users in account/balance model-based sharded blockchains and identify a phenomenon in which accounts (users' wallets and smart contracts) eventually get concentrated in a few shards, making shard loads unfair. This phenomenon leads to bad user experiences, such as delays in transaction inclusions and increased transaction fees. To solve this problem, we propose two load balancing methods in account/balance model-based sharded blockchains. Both methods perform load balancing by periodically reassigning accounts: in the first method, the blockchain protocol itself performs load balancing and in the second method, wallets perform load balancing. We discuss the pros and cons of the two protocols, and apply the protocols to the execution sharding in Ethereum 2.0, an existing sharding design. Further, we analyze by simulation how the protocols behave to confirm that we can observe smaller transaction delays and fees. As a result, we released the simulation program as “Shargri-La,” a simulator designed for general-purpose user behavior analysis on the execution sharding in Ethereum 2.0.

Image resolution under the effect of color speckle was successfully measured for a raster-scan mobile projector, using the modified contrast modulation method. This method was based on the eye-diagram analysis for distinguishing the binary image signals, black-and-white line pairs. The image resolution and the related metrics, illuminance, chromaticity, and speckle contrast were measured at the nine regions on the full-frame area projected on a standard diffusive reflectance screen. The nonuniformity data over the nine regions were discussed and analyzed.