The intrusion detection method of power industrial control systems is a crucial aspect assuring security. However, traditional methods have two drawbacks: first, they are mainly used for defending information and lack the ability to detect attacks against systems; second, although machine learning-based perform well with default hyperparameters, optimizing hyperparameters can significantly improve its performance. In response these limitations, random forest (RF)-based model proposed....

Abstract Bacillus thuringiensis (Bt), one of the most successful biopesticides, may expand its potential by producing bacteriocins (thuricins). The aim this study was to investigate antimicrobial a novel Bt bacteriocin, thuricin CspB, produced BRC-ZYR2. results showed that bacteriocin has high similarity with cold-shock protein B (CspB). CspB lost activity after proteinase K treatment; however it active at 60 °C for 30 min and stable in pH range 5–7. partial loss treatments lipase II...

Abstract The quality parameters of a new type instant rice with higher moisture content and lower water activity (a w) were investigated. product was obtained after combining three types drying processes as follows: air drying, microwave osmotic process. new-type evaluated by comparing it freeze (FD), hot-air (AD), freshly cooked rice. Several discussed, including color, texture, molecule mobility, microstructure, flavor, sensory. comparison showed that the yellowness, hardness, chewiness...

Abstract Randomness is an important research topic in the field of information security, especially cryptography. test techniques are used to examine quality random numbers so that they meet requirements application. The randomness cryptographic algorithms one key concerns algorithm design. People have put forward different standards and for algorithms, as well developed corresponding kits. This paper analyzes technologies general methods, compares them on this basis. At end, actual...

Bacillus thuringiensis (Bt) is one of the most widely used and studied biopesticides. However, it vulnerable to influence ultraviolet (UV) radiation, causing shorter persistence under field conditions. To obtain a high-active effective Bt new product, main objective this study highly UV-resistant mutant from mosquitocidal LLP29 through UV exposure. After 19 rounds exposure, named LLP29-M19 was obtained, showing resistance radiation for up 67 min. The fatality rate 95%, which slightly higher...

In recent years, neural networks based on attention mechanisms have been increasingly widely used in speech recognition, separation, enhancement, and other fields. particular, the convolution-augmented transformer has achieved good performance as it can combine advantages of convolution self-attention. Recently, gated unit (GAU) proposed. Compared with traditional multi-head self-attention, approaches GAU are effective computationally efficient. this article, we propose a network for...

Automatic generation of painting images is an interesting and difficult task, especially for regional traditional paintings with unique cultural styles while lacking large-scale training sets. In this paper, a hierarchical method proposed, which can disentangle the content style. By mimicking human process, proposed introduces multiple blocks first gradually generates image contents. each block, spatial self-modulation module to inject local details preserving global layout. After...

Shape modelling is very important in many tasks of computer vision the internet things. Boltzmann Machine (SBM) a strong shape model, having ability to capture details object by introducing Local Receptive Fields (LRF) and weight sharing into deep learning architecture. However, applying LRF only single layer restrict its capabilities more de-tails representation local parts. In this paper, we propose new model based on Deep (DBM) which call Multi-Scale (MSSBM). By hierarchically...

Abstract In response to the background penetration problem of unsupervised style transfer algorithms in most cases, a Transformer network DualGGAN based on dual generators and fusion relative position encoding is proposed. The trained using least squares generative adversarial network, neural used as image feature extractor generate maps obtain facial features with attention weights from maps, utilizing mask loss jointly constrain region transfer. experimental results show that effectively...

In this paper, our group came up with a new method that automatically creates knowledge graphs from the natural language. It combines different technologies, such as parse of sentence, Resources Description Framework (RDF), and NoSQL database. For sentences analysis part, we used NLTK Stanford coreNLP to get trees language sentences. Then extract RDF triplets trees, which can be create graphs. Test compare them on Visual Genome datasets University because they include many images regions...

Autonomous driving vehicles (ADVs) are sleeping giant intelligent machines that perceive their environment and make decisions. Most existing ADSs built as hand-engineered perception-planning-control pipelines. However, designing generalized handcrafted rules for autonomous in an urban is complex. An alternative approach imitation learning (IL) from human demonstrations. most previous studies on IL face several critical challenges: (1) poor generalization ability toward the unseen due to...

The traditional encryption algorithms such as DES, AES are considered too slow for image encryption, which cannot meet the real-time requirements. Therefore, this paper proposes a parallel algorithm based on cellular automaton and chaotic map. In scrambling stage, keystream generated by 2D-LASM is used to scramble rows columns of in parallel. diffusion scrambled constructed into two identical square bit matrices initial state balance rule, encrypted obtained iterating multiple times. sub key...

Background subtraction algorithm is essential for video processing. such as target tracing, gesture recognition and gait recognition. ViBe has been widely used because of easy implementation high efficiency. However, the would produce a ghost imaging when speed moving changes. On other hand, challenged to adapt change environment by misjudge shadow foreground target. Moreover, it also inability handle well interference caused camera jitter. Aiming deficiencies ViBe, we propose new Gc_IViBe,...

With the rapid development of convolutional neural networks, performance face detection algorithms has been improved. However, recall rate lightweight algorithm is low when it applied to small-scale and dense face. And high-speed high-accuracy should be considered at same time. To solve problems, we propose a based on YOLOv3 key point. The contributions include: (1) A algorithm, YOLOFKP, detector, with scale invariance enlarged receptive field, proposed YOLOv3. (2) Use mosaic for data...

摘要 利用差示扫描量热分析仪(DSC)和淀粉快速黏度分析仪(RVA)测定了14个高粱(Sorghum bicolor)品系的直链淀粉含 量和淀粉糊化特性参数, 并分析了各特征参数的变化规律及其相互关系。研究结果表明, 高粱品系间直链淀粉含量存在较 大差异, 变幅为0

<p indent="0mm">Currently, most image inpainting approaches focus on improving the quality of inpainted images while ignoring requirement for a low-end platform. In this work, we propose lightweight group convolution module in place vanilla convolution. This module, named attention-based (LABGC), uses depth-wise to achieve multi-scale inpainting. Meanwhile, alleviate information constraint between channels, cycle channel attention mechanism. Then our proposed LRGAN network, use LABGC. We...

Purpose How to model blind image deblurring that arises when a camera undergoes ego-motion while observing static and close scene. In particular, this paper aims detail how the blurry can be restored under sequence of linear point spread function (PSF) are derived from 6-degree freedom (DOF) camera’s accurate path during long exposure time. Design/methodology/approach There two existing techniques, namely, an estimation PSF deconvolution. Based on online short-period inertial measurement...