FinScan: Empowering Fish Consumers using Accurate Prediction

Fish fraud harms consumers & ecosystems, mislabeling species extends to aquarium fish, harming diversity. Combatting it is vital for biodiversity & safety through accurate labeling & awareness.

Comment

Inspiration

One day while purchasing a fish species from the local market,we were not able to clearly identify the tyoe of fish we were interested to purchase.The shopkeeper on the other hand scammed us by providing a similar looking species at a higher price.Thus,we thought about the need of correctly identifying edible fish species.For that matter,any fish species(aquarium based fish also included).

What it does

With the help of a mobile based application,the user can correctly identify the fish with the help of a single click.

How we built it

For building this application,we manually created a dataset by clicking pictures of fish species from various local sources including aquariums and fish markets.Then,with the the help of Machine Learning algorithms we were able to correctly extract the features required for identifying a fish namely fins,tail and body etc.

Challenges we ran into

Data Integration,Computational time,Hyper-parameter Tuning,Overfitting,Classifying non-fish images

Accomplishments that we're proud of

Successfully tried in local fish market and got approved by a domain expert in the Fish breeding industry.

What we learned

Handling image data

What's next for FinScan: Empowering Fish Consumers using Accurate Prediction

Includes adding a chatbot feature for interacting and adding extra information about the fish species and also improving the overall GUI of the App.

Built With

Share this project:

Updates