Today, half of the world relies on fish as a major source of animal protein, but nearly 90% of our wild fishing stocks have depleted. With the growing population, we need to expand ocean farming. However, fish farmers don't have the tools and that much insight into the health of their fish; this leads to outbreaks of disease and pollution, which are costly for farmers and detrimental to the environment.
When I came across this problem, I was SO SHOCKED by the quantitative data on this and decided to look into how I could make an impact. Interestingly enough, we can work towards this by leveraging advancements in the computer vision space, especially with image classification, segmentation, and other analysis techniques.
As a result, I read a research paper about these computer vision techniques and am currently working on 2 projects:
- fish classification project to classify 9 classes of seafood
- fish segmentation project to segment out where a fish is in the image