Hi, I am Harshit, a third-year ECE major at Indian Institute of Technology, Roorkee. I was interested in the project "Computer Vision Based PPI Tool". I had some doubts in my mind regarding the idea: I was thinking of the following protoype: 1.) The user would enter the app, and upload some images. 2.) He would be given the option to select the country, for which he wishes to calculate the PPI. (*Doubt*: How many countries to provide the support for?) 3.) Next, the questions of the specific country would appear with the confident answers to those questions.(e.g. Is refrigerator present?, Is gas stove present?) Also some questions would be unanswered, like the education of the lead female member(*Doubt*: Is there a way we can answer such questions?) 4.) The user would be able to manually correct any answer, and then the final PPI would be calculated. I request you to suggest any improvements/changes in it. Secondly, regarding the training on Google Cloud Vision, how would I get/prepare the training set?
Thanks in advance. Regards, Harshit
