The aim of this paper is to propose effective attentional regions for fine-grained visual recognition. Based on the Spatial Transformers' capability of spatial manipulation within networks, we propose an extension model, the Attention-Guided Spatial Transformer Networks (AG-STNs). This model can guide the Spatial Transformers with hard-coded attentional regions at first. Then such guidance can be turned off, and the network model will adjust the region learning in terms of the location and scale. Such adjustment is conditioned to the classification loss so that it is actually optimized for better recognition results. With this model, we are able to successfully capture detailed attentional information. Also, the AG-STNs are able to capture attentional information in multiple levels, and different levels of attentional information are complementary to each other in our experiments. A fusion of them brings better results.

In this letter, we propose a novel framework to estimate the joint distribution of multiple Local Binary Patterns (LBPs). Multiple LBPs extracted from the same central pixel are first encoded using handcrafted encoding schemes to achieve rotation invariance, and the outputs are further encoded through a pre-trained Restricted Boltzmann Machine (RBM) to reduce the dimension of features. RBM has been successfully used as binary feature detectors and the binary-valued units of RBM seamlessly adapt to LBP. The proposed feature is called RBM-LBP. Experiments on the CUReT and Outex databases show that RBM-LBP is superior to conventional handcrafted encodings and more powerful in estimating the joint distribution of multiple LBPs.

Deep Graphical Model (DGM) based on Generative Adversarial Nets (GANs) has shown promise in image generation and latent variable inference. One of the typical models is the Iterative Adversarial Inference model (GibbsNet), which learns the joint distribution between the data and its latent variable. We present RGNet (Re-inference GibbsNet) which introduces a re-inference chain in GibbsNet to improve the quality of generated samples and inferred latent variables. RGNet consists of the generative, inference, and discriminative networks. An adversarial game is cast between the generative and inference networks and the discriminative network. The discriminative network is trained to distinguish between (i) the joint inference-latent/data-space pairs and re-inference-latent/data-space pairs and (ii) the joint sampled-latent/generated-data-space pairs. We show empirically that RGNet surpasses GibbsNet in the quality of inferred latent variables and achieves comparable performance on image generation and inpainting tasks.

In this letter, we propose a novel texture descriptor that takes advantage of an anisotropic neighborhood. A brand new encoding scheme called Reflection and Rotation Invariant Uniform Patterns (rriu2) is proposed to explore local structures of textures. The proposed descriptor is called Oriented Local Binary Patterns (OLBP). OLBP may be incorporated into other varieties of Local Binary Patterns (LBP) to obtain more powerful texture descriptors. Experimental results on CUReT and Outex databases show that OLBP not only significantly outperforms LBP, but also demonstrates great robustness to rotation and illuminant changes.

Fine-grained image classification is a difficult problem, and previous studies mainly overcome this problem by locating multiple discriminative regions in different scales and then aggregating complementary information explored from the located regions. However, locating discriminative regions introduces heavy overhead and is not suitable for real-world application. In this paper, we propose the recursive multi-scale channel-spatial attention module (RMCSAM) for addressing this problem. Following the experience of previous research on fine-grained image classification, RMCSAM explores multi-scale attentional information. However, the attentional information is explored by recursively refining the deep feature maps of a convolutional neural network (CNN) to better correspond to multi-scale channel-wise and spatial-wise attention, instead of localizing attention regions. In this way, RMCSAM provides a lightweight module that can be inserted into standard CNNs. Experimental results show that RMCSAM can improve the classification accuracy and attention capturing ability over baselines. Also, RMCSAM performs better than other state-of-the-art attention modules in fine-grained image classification, and is complementary to some state-of-the-art approaches for fine-grained image classification. Code is available at https://github.com/Dichao-Liu/Recursive-Multi-Scale-Channel-Spatial-Attention-Module.

Genetic algorithms have been shown to be very useful in a variety of search and optimization problems. In this paper we present a genetic algorithm for maximum independent set problem. We adopt a permutation encoding with a greedy decoding to solve the problem. The DIMACS benchmark graphs are used to test our algorithm. For most graphs solutions found by our algorithm are optimal, and there are also a few exceptions that solutions found by the algorithm are almost as large as maximum clique sizes. We also compare our algorithm with a hybrid genetic algorithm, called GMCA, and one of the best existing maximum clique algorithms, called CBH. The exiperimental results show that our algorithm outperformed two of the best approaches by GMCA and CBH in final solutions.

Combining scheduled channel polling with channel diversity is a promising way for a MAC protocol to achieve high energy efficiency and performance under both light and heavy traffic conditions. However, the deafness problem may cancel out the benefit of channel diversity. In this paper, we first investigate the deafness problem of scheduled multi-channel polling MACs with experiments. Then we propose and evaluate two schemes to handle the deafness problem. Our experiment shows that deafness is a significant reason for performance degradation in scheduled multi-channel polling MACs. A proper scheme should be chosen depending on the traffic pattern and the design objective.

With the development of power market demand response capability, load aggregators play a more important role in the coordination between power grid and users. They have a wealth of user side business data resources related to user demand, load management and equipment operation. By building a business model of business data resource utilization and innovating the content and mode of intelligent power service, it can guide the friendly interaction between power supply, power grid and load, effectively improve the flexibility of power grid regulation, speed up demand response and refine load management. In view of the current situation of insufficient utilization of business resources, low user participation and imperfect business model, this paper analyzes the process of home appliance enterprises participating in peak shaving and valley filling (PSVF) as load aggregators, and expounds the relationship between the participants in the power market; a business service model of smart home appliance participating in PSVF based on cloud platform is put forward; the market value created by home appliance business resources for each participant under the joint action of market-oriented means, information technology and power consumption technology is discussed, and typical business scenarios are listed; taking Haier business resource analysis as an example, the feasibility of the proposed business model in innovating the content and value realization of intelligent power consumption services is proved.

The privacy of users' data has become a big issue for cloud service. This research focuses on image cloud database and the function of similarity search. To enhance security for such database, we propose a framework of privacy-enhanced search scheme, while all the images in the database are encrypted, and similarity image search is still supported.

An advanced interest point detector is proposed to improve the Hessian-Matrix based detector of the SURF algorithm. Round-like shapes are utilized as the filter shape to calculate of the Hessian determinant. Dxy can be acquired from approximate round areas, while the regions for computing Dyy or Dxx are designed with the consideration to symmetry and a balance of pixel number. Experimental results indicate that the proposed method has higher repeatability than the one used in SURF, especially in the aspects of rotation and viewpoint, due to the centrosymmetry of the proposed filter shapes. The results of image matching also show that more precision can be gained with the application of proposed detector.

Halo doping profile is used in nanoscale surrounding-gate MOSFETs to suppress short channel effect and improve current driving capability. Analytical surface potential and threshold voltage models are derived based on the analytical solution of Poisson's equation for the fully depleted symmetric and asymmetric halo-doped MOSFETs. The validity of the analytical models is verified using 3D numerical simulation. The performance of the halo-doped MOSFETs are studied and compared with the uniformly doped surrounding-gate MOSFETs. It is shown that the halo-doped channel can suppress threshold voltage roll-off and drain-induced barrier lowering, and improve carrier transport efficiency. The asymmetric halo structure is better in suppressing hot carrier effect than the symmetric halo structure.

In this paper, by using the properties of the cyclic Hadamard matrices of order 4t, an infinite class of (4t-1)-variable 2-resilient rotation symmetric Boolean functions is constructed, and the nonlinearity of the constructed functions are also studied. To the best of our knowledge, this is the first class of direct constructions of 2-resilient rotation symmetric Boolean functions. The spirit of this method is different from the known methods depending on the solutions of an equation system proposed by Du Jiao, et al. Several situations are examined, as the direct corollaries, three classes of (4t-1)-variable 2-resilient rotation symmetric Boolean functions are proposed based on the corresponding sequences, such as m sequences, Legendre sequences, and twin primes sequences respectively.

The IEEE 802.11 standard has been extensively deployed all over the world. Many studies have been put on its performance, especially throughput. Most research focused on the analysis of saturated throughput, but non-saturated situation is more usual in practice. By extending a saturation throughput model, a concise and novel model is proposed in this paper, which can be used to analyze both saturated and non-saturated conditions. Moreover, the model can also deal with the heterogeneous condition, which allows stations to have different traffic. Different access mechanisms and packet payloads are used in simulation to validate it, and the results show that the model is accurate.

Real-time and reliable radio communication is essential for wireless control systems (WCS). In WCS, preambles create significant overhead and affect the real-time capability since payloads are typically small. To shorten the preamble transmission time in OFDM systems, previous works have considered adopting either time-direction extrapolation (TDE) or frequency-direction interpolation (FDI) for channel estimation which however result in poor performance in fast fading channels and frequency-selective fading channels, respectively. In this work, we propose a subcarrier-selectable short preamble (SSSP) by introducing selectability to subcarrier sampling patterns of a preamble such that it can provide full sampling coverage of all subcarriers with several preamble transmissions. In addition, we introduce adaptability to a channel estimation algorithm for the SSSP so that it conforms to both fast and frequency-selective channels. Simulation results validate the feasibility of the proposed method in terms of the reliability and real-time capability. In particular, the SSSP scheme shows its advantage in flexibility as it can provide a low error rate and short communication time in various channel conditions.

This paper introduces and uses spider monkey optimization (SMO) for synthesis sparse linear arrays, which are composed of a uniformly spaced core subarray and an extended sparse subarray. The amplitudes of all the elements and the locations of elements in the extended sparse subarray are optimized by the SMO algorithm to reduce the side lobe levels of the whole array, under a set of practical constraints. To show the efficiency of SMO, different examples are presented and solved. Simulation results of the sparse arrays designed by SMO are compared with published results to verify the effectiveness of the SMO method.

Standard neural machine translation (NMT) is on the assumption that the document-level context is independent. Most existing document-level NMT approaches are satisfied with a smattering sense of global document-level information, while this work focuses on exploiting detailed document-level context in terms of a memory network. The capacity of the memory network that detecting the most relevant part of the current sentence from memory renders a natural solution to model the rich document-level context. In this work, the proposed document-aware memory network is implemented to enhance the Transformer NMT baseline. Experiments on several tasks show that the proposed method significantly improves the NMT performance over strong Transformer baselines and other related studies.

We propose a motion detection model, which is suitable for higher speed operation than the video rate, inspired by the neuronal propagation in the hippocampus in the brain. The model detects motion of edges, which are extracted from monocular image sequences, on specified 2D maps without image matching. We introduce gating units into a CA3-CA1 model, where CA3 and CA1 are the names of hippocampal regions. We use the function of gating units to reduce mismatching for applying our model in complicated situations. We also propose a map-division method to achieve accurate detection. We have evaluated the performance of the proposed model by using artificial and real image sequences. The results show that the proposed model can run up to 1.0 ms/frame if using a resolution of 6460 units division of 320240 pixels image. The detection rate of moving edges is achieved about 99% under a complicated situation. We have also verified that the proposed model can achieve accurate detection of approaching objects at high frame rate (>100 fps), which is better than conventional models, provided we can obtain accurate positions of image features and filter out the origins of false positive results in the post-processing.

In this paper, we study partial key exposure attacks on RSA where the number of unexposed blocks of the private key is greater than or equal to one. This situation, called generalized framework of partial key exposure attack, was first shown by Sarkar [22] in 2011. Under a certain condition for the values of exposed bits, we present a new attack which needs fewer exposed bits and thus improves the result in [22]. Our work is a generalization of [28], and the approach is based on Coppersmith's method and the technique of unravelled linearization.

At FSE 2014, Grosso et al. proposed LS-designs which are a family of bitslice ciphers aiming at efficient masked implementations against side-channel analysis. They also presented two specific LS-designs, namely the non-involutive cipher Fantomas and the involutive cipher Robin. The designers claimed that the longest impossible differentials of these two ciphers only span 3 rounds. In this paper, for the two ciphers, we construct 4-round impossible differentials which are one round more than the longest impossible differentials found by the designers. Furthermore, with the 4-round impossible differentials, we propose impossible differential attacks on Fantomas and Robin reduced to 6 rounds (out of the full 12/16 rounds). Both of the attacks need 2119 chosen plaintexts and 2101.81 6-round encryptions.

As the mainstream approach in the field of machine translation, neural machine translation (NMT) has achieved great improvements on many rich-source languages, but performance of NMT for low-resource languages ae not very good yet. This paper uses data enhancement technology to construct Mongolian-Chinese pseudo parallel corpus, so as to improve the translation ability of Mongolian-Chinese translation model. Experiments show that the above methods can improve the translation ability of the translation model. Finally, a translation model trained with large-scale pseudo parallel corpus and integrated with soft context data enhancement technology is obtained, and its BLEU value is 39.3.