Article From:

>, pair information.

  • Pairing members
    031602206 Chen Wen Jie
    031602209 Chen Zhi Wei
  • This homework blog link
  • Pairs of students blog links
  • Prototype design tool: Adobe XD CC
  • pdfFile



According to the topic, users hope they can understand the research hotspots of CVPR, ICCV, ECCV. They need to collect papers from the three top conferences in recent years for screening and analysis, so as to find out the hot research directions, so that they can finish an excellent paper standing at the forefront of the times before graduation.

It is time-consuming and inefficient to find papers from online lists. Therefore, users urgently need a tool that can quickly crawl the basic information of papers, understand the research direction of papers, and analyze the research hotspots of papers.

  • User given list of papers
    • Crawl through the list of papers, title, summary, original link
    • Add and delete the list of papers (this year, nearly two years, nearly three years).
  • The crawling information is structured to analyze TOP10 hot areas or popular research directions.
    • We can filter and analyze paper attributes (oral, spotlight, poster).
    • Form an intuitive way of viewing such as hot word maps.
  • Papers can be retrieved. When the user enters the basic information such as paper number, title, author, etc., the relevant paper, source code, home page and other information can be analyzed and returned.

  • It can compare the thermal trend of hot words between different years and different summit.

  • Data statistics can be carried out, such as the analysis of each country’s employment articles, the analysis of each school’s employment articles, which school’s research direction is stronger, etc.

Above is the user has clearly given the requirements, but to understand the user’s needs can not only rely on the user’s description, but also need the developer to stand in the user’s perspective to think. After some consideration, we believe that we may be able to add a need for collection of papers. When a user sees something he likes very much, maybePapers that will be used in the future can be added to the collection, and then can be found directly in their favorites, more convenient and fast.


Considering the applicability, the system is applied to the Web terminal. In view of the above requirements, the solutions we discussed are as follows.

  • Provide a blank paper list, users can input the paper information they want to crawl, users can operate on the paper list at any time, and automatically jump to the paper information page after the crawl is completed.
  • To analyze the crawled papers information and provide a variety of selection conditions, such as selecting all papers / selected schools / selected countries / selected years / selected papers attributes, analyzing top 10 hot research areas or hot directions, showing them in the form of a hot word map or a broken line mapCome;
  • Set up the search bar, and provide the paper number, title, author, paper attributes and other search conditions, after the search can return the details of the paper. If the selected conditions are the author, then all the papers of the author will be returned after the search.
  • Set the year, top and other screening conditions, return the selected conditions of the hot words trend comparison chart;
  • Users can enter the name of the country/school/author they want to know and return to the country/school/author’s main research directions and fields in recent years.


From the user’s point of view, we believe that the use of this platform for statistical analysis of paper information benefits as follows:

  • It saves users a lot of time to look up information. Nowadays, there are so many papers on the Internet. There are so many kinds of papers. It is impossible for users to browse every paper to see if it is useful for them. Using this platform, users can quickly and accurately obtain the paper information they need, and get the most fullThe most careful statistical analysis can save unnecessary time and improve learning efficiency.
  • In the face of a large number of papers, users often feel unable to find the direction of learning, the use of this platform services, users can be cursory knowledge of the current top conference research direction of academic origin, find the way to read, find useful papers for themselves, improve their academic level;
  • The platform is simple and clear, beautiful and easy to understand.No advertising,Users can conveniently operate according to their own ideas and have a high degree of friendliness to users.


  • In terms of peripheral functions, the interface is beautiful, simple and well designed, and it can give users a very good sense of experience.
  • In terms of the necessary requirements, the paper crawling speed is fast, the selection and analysis of the paper is meticulous and reliable.
  • This platform has a collection function of papers, users can easily collect their favorite papers, friendly degree max.
  • With data statistics and data analysis functions, users can very intuitively view the amount of papers in various countries, schools and the main research directions of each school.
  • At present, CiteSeer and Microsoft academic search platforms on the market also provide paper retrieval functions, but relatively lack of data analysis and statistical functions.

” >(5)Delivery–Extension

  • Since the core function of this platform is to crawl papers and analyze hot research directions, so the main user groups should be senior students, graduate students and teachers. We will first develop our personal relationships, introduce our products to others through our classmates and friends, and then promote them in our college.Wide open.
  • After opening up the market in our college, we should strive for the support of the school-level Department associations or school management departments, and then promote it to the whole school to let more students and teachers know and use our products.
  • In the use of the process of users to investigate, collect feedback from users, analyze the defects and shortcomings of the product, constantly improve product quality, enhance user experience, in the user community for a better reputation.

three, pairs of photos




PSP2.1Personal Software Process StagesEstimated time (minutes)Actual time consumption (minutes)Planningplan6060· Estimate· Estimate how much time this task will take.6060DevelopmentDevelopment19201620· Analysis· Requirements analysis (including learning new technology)120240· Design Spec· Generate design documents12060· Design Review· design review6060· Coding Standard· Code specification (to develop appropriate specifications for current development)00· Design· Specific design120240· Coding· Specific coding1440960· Code Review· Code Review6060· Test· Test (self test, modify code, submit Revision)00ReportingPresentation130130· Test Repor· Test report00· Size Measurement· Computational effort1010· Postmortem & Process Improvement Plan· Summarize afterwards and propose process improvement plan.120120Total21101810

Week NNew code lineCumulative code lineLearning time (hours) this weekCumulative learning time (hours)Important growth12002001515Learn the use of Java and IDEA.21010Read the method of building, understand the NABCD model, and learn the use of prototype tools.

seven, difficulties encountered and Solutions

  • I spent a lot of time on understanding the topic. Although I asked my sister about some questions, I still spent a lot of time on discussing and understanding the topic.
  • Aesthetic black holes, clearly in mind very good, but the effect is not as good as expected, in beautifying some details encountered greater difficulties.
  • The schedule was not reasonable enough. Both of them had a lot of work to do. The homework they had done last week was not finished until the last day. Next time pairing work, we must pay attention to the time arrangement.

    Post Views: 1