The problem Fish Lens solves

Helps the concerned authorities in identifying the proportion of deiferent species of fishes from day to day harvest.
During peak season the automated model helps in minimizing the errors in manual assesment of the fish harvest/stock.
Helping the fishermen community to identify the the amount they can generate from thier daily harvest via our android application.

Challenges we ran into

Converting the keras model to adaptable tensorflow lite model so that it can be integrated with java.
Integration of the CNN model to java.