Developed various machine learning models in order to classify pictures of fishes from CCTV-like cameras on-board fishing boats. The final model used classical pre-processing, Isomap, and neural network as a classifier.
We had a strong constraint of not having access to a lot of computer ressources, so deep-learning was out of the question. Most of the project was therefore data exploration, creation of good features from the images, and optimisation.
Disclaimer: the image used is not from the project (although it very well could be), since my images were protected by a Non-Discolure Agreement.