4th Tellus Satellite Challenge using AI MLcoreview
We made it to the top 10% !!
The evaluation criteria were the accuracy of predicting the coastline from a satellite image using AI ML techniques. This was measured by the overlap of the coastline predicted by our algorithm over the actual coastline. The more they overlap less is the error score. We started with an error score of 110 and brought it down to 34, which means we could detect the coastline with approximately 80% accuracy.
We made it to the top 10% out of the 800 participants. ?Unfortunately, we got to know about the competition a couple of weeks after the actual start date. Yet we achieved great success in just 45 days of a team effort.