The goal of this project is to create labeled dataset, which will enable training of neural network detecting floating objects. This technology could be used for finding and saving people lost at sea.
The dataset will be made publicaly available for free download for other programmers from around the world to use in their work.
Creation of a good dataset is the key to success for every project linked with computer vision, neural networks and artifficial intelligence. For correct detecion of objects on photos or videos computer needs to first analyze tens of thousands of similiar pictures. Unfortunately, despite collecting some of the materials, I still need more data to train my algorithm for it to successfully detect objects in different surroundings. If you have any videos or photos of objects floating on the water surface made in “bird’s eye view”, I would very much appreciate sending them to me.
Few minutes of video is enough
In machine learning the most important thing is the diversity of data used in training of the algorithm. That’s why even a couple of minutes of Your video can make a big difference in precision and quality of my algorithm.
Join partners of the project by sending videos or photos.