Jiameng’s Status Update for 02/29/2020

Accomplishments this week

• Try cropping out the title part of images and use MSE to compare
  • Does not work because the accuracy is lower than expected for comparing the titles due to positioning and other reasons
• Try using Tesseract to find the text in the music score
  • Successfully transformed most of the word on the music score and printed those out
  • Also try to crop out only the title part from the music score and figure out what the text is to avoid getting extra texts
    • Failed because for different music scores, it is hard to get exactly the title part and will sometimes get other text to be cropped or incomplete titles
  • Try to determine where the text is based on the largest text on the page (usually the case because title should be larger than other texts)
    • Failed because for a music score, the score part is considered a large piece of text (actually the contour algorithm is looking for large chunks of rectangles, which works if the page is filled with texts) and therefore the title is ignored by the algorithm
  • Try to determine where the text is by taking the uppermost text on the page, and deciding if the next contour belongs to the title part based on its difference in height from the previous contour
    • Works well because usually there’s no other text above the title for a music score, if we do not test with extreme cases (like taking a picture of the music score with a large poster with a big title behind and above the music score), also, the algorithm requires some resolution of the picture, otherwise it could not recognize the text, so it works well with the title of music score which is close to the camera and could ignore what’s behind the music score in most cases
• Created a web application using Django with a basic login page and registration page
  • Features completed includes:
    • Added a database for the login form and registration form
    • Added the CSS files for the application (only for login page and registration page)
    • Build connection between login page and registration page
    • Loaded the music scores into the web application database (which is not deployed yet and is stored locally for  the moment)

Progress for schedule

• On schedule

Deliverables I hope to accomplish next week

• Get the website setup and working with basic features like going from one site to another with hyperlinks
• Try the library cadenCV and recognize all the music notes on a music score (don’t need to recognize repeat position yet)