Siamese visual tracking with multilayer feature fusion and corner distance IoU loss

The tracker based on the Siamese network regards tracking tasks as solving a similarity problem between the target template and search area. Using shallow networks and offline training, these trackers perform well in simple scenarios. However, due to the lack of semantic information, they have difficulty meeting the accuracy requirements of the task when faced with complex backgrounds and other challenging scenarios. In response to this problem, we propose a new model, which uses the improved ResNet-22 network to extract deep features with more semantic information. Multilayer feature fusion is used to obtain a high-quality score map to reduce the influence of interference factors in the complex background on the tracker. In addition, we propose a more powerful Corner Distance IoU (intersection over union) loss function so that the algorithm can better regression to the bounding box. In the experiments, the tracker was extensively evaluated on the object tracking benchmark data sets, OTB2013 and OTB2015, and the visual object tracking data sets, VOT2016 and VOT2017, and achieved competitive performance, proving the effectiveness of this method.     

Wearable computers as a virtual environment interface for people with visual impairment

People who are totally blind or who have vevere visual impairments (e.g. less than 20/200 acuity, central macular scotomas, or advanced diabetic retinopathy) see the environment in a fashion that may be completely foreign to those who operate in a very visual fashion. For this population, the visual complexity of the environment is not a concern. What is of concern are salient features found within the environment that relate to their ability to navigate successfully in, and/or interact with, the environment as needed. Toward the purpose of representing these salient features in comprehensive form, investigators at the Atlanta Veterans Affairs Research and Development Center are employing wearable computer technology to develop a virtual environment interface. The long-range goal is to create a simplistic virtual representation of the environment that includes only features related to the current navigational task and/or interactive needs of the person. In a completed study, the use of digital infrared transmitters as beacons representing salient features of the environment was explored. The purpose of a current study now in progress is to evaluate and compare various user interface structures that were suggested by subjects during the preliminary study. The problem of interest in the current study is street-crossing; however, the results of this research should be applicable to many other problems, including identifying and locating building entrances, and identifying, locating and interacting with electronic devices such as information kiosks, automated teller machines, and self-serve point-of-sale terminals. The long-range result desired is a wearable computer with which one can easily identify and interact with a wide variety of devices in the environment via one familiar, easy-to-use interface.     

GOST 34.10—A brief overview of Russia's DSA

GOST 34.10 is Russia's DSA. Like its US counterpart, GOST is an ElGamal-like signature scheme used in Schnorr mode. It is similar to NIST DSA in many aspects. In this paper we will overview GOST 34.10 and discuss the three main differences between the two algorithms, (i) GOST's principal design criterion does not seem to be computational efficiency: the algorithm is 1.6 times slower than the DSA and produces 512-bit signatures. This is mainly due to the usage of the modulus q which is at least 254 bits long. During verification, modular inverses are computed by exponentiation (while the Extended Euclidian algorithm is roughly 100 times faster for this parameter size) and the generation of the public parameters is much more complicated than in the DSA. This choice of the parameters makes GOST 34.10 very secure. (ii) GOST signers do not have to generate modular inverses as the basic signature equation is s = xr + mk (mod q) instead of (mod q). (iii) GOST's hash function (the Russian equivalent of the SHA) is the standard GOST 34.11 which uses the block cipher GOST 28147 (partially classified) as a building block. The hash function will be briefly described. Copyright     

Distinguishing attacks on stream ciphers based on arrays of pseudo-random words

In numerous modern stream ciphers, the internal state consists of a large array of pseudo-random words, while the output key-stream is a relatively simple function of the state. It has been heuristically shown in several situations [3], [8], [9], [10], [11] and [14] that this structure may lead to distinguishing attacks on the cipher. In this note we present a more rigorous treatment of this structural attack. First, we present a rigorous proof of the main probabilistic claim behind it in the basic cases. We then apply it concretely to the cipher sn3 [12], and demonstrate that the heuristic assumptions of the attack are remarkably precise in more complicated cases.     

Cryptanalysis of RSA with two decryption exponents

In this paper, we consider RSA with N=pq, where p,q are of same bit size, i.e., q<p<2q. We study the weaknesses of RSA when multiple encryption and decryption exponents are considered with same RSA modulus N. A decade back, Howgrave-Graham and Seifert (CQRE 1999) studied this problem in detail and presented the bounds on the decryption exponents for which RSA is weak. For the case of two decryption exponents, the bound was N^0.357. We have exploited a different lattice based technique to show that RSA is weak beyond this bound. Our analysis provides improved results and it shows that for two exponents, RSA is weak when the RSA decryption exponents are less than N^0.416. Moreover, we get further improvement in the bound when some of the most significant bits (MSBs) of the decryption exponents are same (but unknown).     

A high quality and small shadow size visual secret sharing scheme based on hybrid strategy for grayscale images

Many secret sharing schemes for digital images have been developed in recent decades. Traditional schemes typically must deal with the problem of computational complexity, and other visual secret sharing schemes come with a higher transmission cost and storage cost; that is, each shadow size is m times as big as the original secret image. The new (2,n) secret sharing scheme for grayscale images proposed in this paper is based a combination of acceptable image quality using block truncation coding (BTC), high compression ratio discrete wavelet transform (DWT) and good subjective performance of the vector quantization (VQ) technique. Experimental results confirm that our proposed scheme not only generates a high quality reconstructed original image but also generates small, random-like grayscale shadows.     

Fixed points of the RSA encryption algorithm

In this paper the problem of the number of fixed points for an RSA algorithm is considered. This is an important question from the point of view of any cryptosystem. We have estimated the expected value of this number for randomly chosen RSA parameters. It turned out that it is O(ln²n), and the probability of finding such a point is O(ln²n/n). Thus, these values are really negligible, which had been intuitively expected.     

应用型本科经管类专业Visual FoxPro程序设计课程教学改革探索

针对目前多数应用型本科院校经济、管理类专业计算机公共基础课Visual FoxPro程序设计教学中存在的问题,本文提出以项目导向、模块驱动的方式来组织教学内容,课堂教学和项目小组协同学习相互促进的教学方式,课内实验和实战演练相结合的训练方法,并给出具体实施办法,实践证明这种教学模式行之有效、教学效果良好。     

Visual Foxpro结构化程序设计教学方法和技巧探讨

通过对《Visual Foxpro程序设计》课程章节教学方法和手段问题进行探讨,激发非计算机专业学生利用Visual Foxpro（以下简称VF）语言熟练编写程序的能力,拓展他们的逻辑思维,技巧转换意识,培养独立思考问题、分析问题和解决问题的能力,并养成良好编写程序代码的风格。