Significant risks for the project we managed this week:
-
Using natural language processing for finding the title of music score input and matching it with database might have two risks in the future
-
For music scores that are not in the database, it will be harder to match because for the same name, there may be multiple different music scores online and even if the name matches, the score may not match, and also, there exists scores without a title, which could be troubling
-
To do natural language processing, and for the current algorithm we have, it requires the photo taken to be relatively clear, and the title text must be higher than all the other texts in the on the music score for the title to be actually recognized as the title of the music score, which is true in most cases
-
Changes made to the existing design of the system:
-
Changes the original decision for doing image matching to recognize which music score in database the input file correspond to, to doing natural language processing on the title of the input music score
-
Changes the original decision for using three stepper motors to using two stepper motors and one servo, because servo is lighter and requires less torque from the stepper motor.
An updated schedule if changes have occurred:
-
Decide to postpone the development of a Django website (which was scheduled for this week) since it is not the most challenging part of our tasks. We moved the development of the MIDI matching algorithm v2 (one that matched slices that are slightly different from each other) forward and plan to start building our website next week.
-
Decide to postpone control the stepper motor with Raspberry Pi. The reason behind is that the stepper motor has just arrived on 2/21, the servo hasn’t arrived yet, and the Raspberry Pi is still missing some cables.
0 Comments