on

FishScan is an app to prevent seafood frauds.

Unfortunately this frauds are very common, and they include:

- Substituting one species for another without changing the label to fraud the customer

- Selling frozen fish as fresh fish

- Selling as fresh fish that are being chemically treated to prevent color changing

Oceana, a marine conservation nonprofit with a recent history of studying seafood mislabeling, published a new report on the state of seafood fraud in the U.S.

They found that 20 percent of the 449 fish they tested were incorrectly labeled. Orders of sea bass were often replaced by giant perch, Alaskan halibut by Greenland turbot, and Florida snapper by lavender jobfish, to name a few.

Oceana made headlines in 2016 by publishing a report finding massive seafood fraud on a global scale. Since then, NOAA created the Seafood Import Monitoring Program (SIMP), to track 13 species deemed at high risk of being fraudulently sold or sourced illegally.

If you've been working in this industry for years is easy to detect and prevent this fraud, if you know where to look is easier to actually detect which fish is that and wether is fresh or not. Unfortunately usually people don't know what to look for and ho to prevent to being scammed. 

Here there is FishScan. FishScan is a mobile application that through AI can easily detect the kind and freshness level of a fish from a picture. In this way the user can simply take a picture on what is buying and be sure that he's paying for the right product. Moreover he can makes sure that the fish is in the state declared on the label (fresh, frozen, etc). This application use Computer Vision and specifically Convolutional Neural Networks and being trained on thousands of images has an accurati rate of 95%. 

This system will prevent vendors to scam customers and to exploit their nescience to increase their profit. On the long run the vendor will stop buying low cost and unethically fished seafood because they won't be able to sell them as fresh fish. 

The process is really easy, the user just need to download the application and to take a picture of the fish. There will be also a quick button to report wether a vendor is trying to sell a fish with a wrong label to try to suppress this unethical approach that unfortunately is really common world wide. 

The project is really scalable and the AI model has already been trained on more than 20 species, but is still in a work in progress version so we can only share the mockup.

The same application will also be released for the amateur fishers that want to see which fish they've just caught.

Usually in fact they are not sure about the species and the minimum size they can catch. In this way fishers will be more responsible and will double check wether the fish just caught can be eaten (if is not an endangered species and if is in the right period of the year and over the minimum size). 

https://drive.google.com/file/d/1RXy5xmoUpHvKJ0Fl0DEH1HELX-M5waKr/view?usp=sharing