2019 International Russian Automation Conference (RusAutoCon)
The authors consider the problem of automatic detection of private scanned documents based on text recognition with deep neural networks. The paper suggests implementing a two-phase approach with the first stage which includes efficient EAST text detection and recognition using Tesseract OCR Engine. Secondly, the authors classify the privacy of a scanned document by deep neural networks applied to the extracted text. After that, a special dataset is gathered in order to train these networks. The experiments show that using OCR Engine for both text detection and segmentation ends up with relatively poor identification of private documents when compared to preliminary text detection with EAST method. Moreover, conventional keyword spotting using the list of sensitive words is less accurate when compared to neural network-based methods. Finally, it was demonstrated that the classification of a bag of most frequent words outperforms traditional text classification techniques with LSTM and convolutional networks.
This paper presents a brief description of a method of automated parametric and structural optimization of temperature control systems of electronic equipment using criteria of solution effectiveness evaluation. Methods and models, that allow automating the process of choosing both thermal regulation (temperature control) tools for electronic equipment and their parameters, are proposed. The problems of formalization of temperature control means selection criteria for electronic components through the cost coefficients, specified by the expert, are considered. The criteria for optimal design allow estimating the effectiveness of heat balance diagram and temperature control means, as well as their parameters. The optimization algorithm, proposed in the article, considers a variety of different continuous states of variable heat balance diagram that depends on the thermal model temperature control means and their parameters. At the same time, temperature control means, that have the greatest impact on the thermal model at the next step with taking into account the cost function, are selected. As an example, the optimal synthesis of temperature control means of a typical electronic equipment block with taking into account the necessity to assess the feasibility of using aluminum heat sinks with electronic components as compared to heat pipes was made.
When measuring small, slowly changing physical parameters, such as parameters of the velocity of low-density gas flows, small values of aerodynamic forces, aerodynamic forces acting on aircraft in conditions of a rarefied gas environment, electrical signals coming from primary converters often have a low level, sometimes reaching very small values, is much less than the level of noise and interference. The most acceptable way of measuring such physical quantities in obtaining reliable results is the electromechanical modulation (interruption) of the oncoming gas flow by means of flow interrupters, which allows obtaining the required higher amplification of the signal of the modulated signal, it is easy to separate the modulated signal from the unmodulated noise and interference signal, even if the level of the useful signal is lower than the level of noise and interference. Successful in solving such a problem is the use of signal accumulation, which increases the signal-to-noise ratio at the converter output. The paper presents the results of an investigation of a high-speed pressure transducer for a gas medium of low density. The stability of the metrological characteristics of the converter is investigated using the method of probabilistic study of the stability of the output characteristic of a membrane-capacitive converter to determine the optimum frequency of electromechanical transformation and to ensure the specified accuracy and stability of its metrological characteristics, taking into account the random nature of the deviation of structural and electrophysical parameters associated with destabilizing factors and technological factors.
The paper presents the implementation of a dynamic routing algorithm intended for use in networks-on-chip with a circulant topology with three generatrices of type C(N; s1, s2, s3) for finding the shortest routes between any two network nodes. The algorithm can be implemented as a RTL state machine in routers for NoCs. The proposed algorithm was tested on sets of optimal circulants. Compared with the classical algorithms A* or Dijkstra, the proposed algorithm does not require to calculate the entire path of the packet, but calculates the port number to which the packet should be sent so that it can reach the destination node. This makes it possible to significantly simplify the structure of the NoC router.