Automatic Marking System For Information Technology Based On Content And Image Analysis

Automatic marking of information technology is based on the documents submitted by the students, access to the document corresponding to the operating point of the attribute value and the standard answer to determine whether the correct operation of the student. Through understanding knowledge demand on information technology textbook, information technology examination outline, 2016 information technology questions of in-depth analysis, Through information technology textbooks IT syllabus, 2016 IT exam in-depth analysis to understand the needs of knowledge, on this basis, developed a set of IT exams and training system, developed a set of information technology exam and practice systems.In this paper, information technology of automatic marking depth study on the shortcomings of the conventional automatic marking technology has been optimized and improved, combined with the potential value of the image entropy achieved sketch recognition. Specific studies in the following areas:(1) In the demand of Taiyuan Chi Lin Information Technology Co., Ltd. are analyzed, and on the basis of information technology teaching material, information technology examination outline, 2016 information technology questions of in-depth study, summarizes the information technology to realize the automatic marking system demand, demand of knowledge points, as well as in the realization of the basic techniques used in the process and the key technology.(2) Focus on the word, excel, ppt document structure analysis, combined with knowledge of the needs analysis and design corresponding to the XML standard answer exam, an operating point of the requested property attribute corresponding score in the exam answers. And XML standard answer exam to further optimize the operation requires no attribute nodes removed, combined with C # Dictionary to achieve optimization in mind. And study the discriminant function execution process, the corresponding subroutine. Introduces several common questions selection algorithm to analyze the strengths and weaknesses of these algorithms, algorithm performance and use of the environment is described in detail in order to choose the most suitable for use herein topics algorithms. And the appropriate knowledge difficulty coefficient are briefly described.(3) The existing design schemes of the Photoshop automatic classification system are studied,, and the inadequacies of existing programs be optimized. On the basis of photoshop document structure on the analysis, detailing the structure of the document object model and the corresponding attributes, design photoshop XML standard answer exam. This article is designed layers positioning algorithm, which can effectively avoid the increase or remove layers during the layer index variation caused by misjudgment, to increase the accuracy graders. Also designed to be a unique sequence matching algorithm, the function execution processes. Describes photoshop automatic identification technology used in the process Jamstyles total integration of information entropy potential of hand-drawn sketches and image characteristics described method, described in detail Jamstyles function and read back the data structure, we proposed to use Jamstyles solve simple use COM not discriminating layer style methodological problems; analysis of the methods commonly used in the sketch retrieval and use of the environment and the advantages and disadvantages, and proposed a total integration of information entropy potential of hand-drawn sketches image features used in this article based on the description method can not address the use of COM sketch discrimination issues.(4) Analysis of the flash document structure, combined with flash knowledge needs analysis and design of the flash XML standard exam. It describes the use of flash in jsfl language for XML to load and read, read jsfl use flash document properties and objects. By the presence of operational requirements document property value and flash flash XML jsfl read attribute node values are compared for automatic grading. Finally, we study the optimization process in the XML standard exam optimization and the use of XML jsfl exam queries.(5) A detailed design of the above algorithms and systems based on C # language used herein, VS2010 programming tools to achieve the commissioning of interface design and program, after extensive testing, to achieve high accuracy in terms of automatic scoring, Real-time is also very good, since the beginning of the promotion in the community has been well received.