Classifies VITB ID cards into hosteller, day scholar, and faculty1.
- Dataset acquired of 11 IDs in different angles and lighting conditions totalling 100 images + 12 control.
- Training was done on 100 images using teachable machine. It took ~2 minutes.
- Used Tensorflow.js to import the model on the webpage, which is just a simple HTML file.
demo.mp4
Footnotes
-
the model doesn't detect faculty ID cards as we didn't have a faculty cards dataset. if you have the means of acquiring these images, feel free to submit a PR. ↩