The PPI (Poverty Probability Index) is a powerful tool to measure poverty index and also a yard stick to measure the progress and impact of poverty outreach programs. To reduce the amount of human intervention in surveying for more accurate and quicker results, I have submitted a proposal in which I plan to make an Android application which will use Machine Learning, Computer Vision through the Google Cloud Vision API and Tensorflow library to perform image analysis and aid the surveying process. The app will render the relevant PPI using the Apache-Fineract SPM API. Further, images may be uploaded to the Cloud Vision Models or fed to the embedded Tensorflow models which are made available using Firebase ML Kit to obtain a report of various objects detected and inferences derived relevant to PPI questionnaire.



Apoorva M K


  • Ishan Khanna
  • Nayan Ambali