3 ms·
It's a nice project, from start to end. Gives you real insight how "real world" ML-backed systems work. Data acquisition, feature engineering, model design / tu
by TrackerFF 4y ago
It's a nice project, from start to end. Gives you real insight how "real world" ML-backed systems work. Data acquisition, feature engineering, model design / tuning, etc.
This kind of system is what you'd expect to see in maybe a graduate level ML/DL class, maybe even Bachelors thesis.
Whether or not it actually works for catching poachers isn't all that important IMO, even though they're marketing it that way. As many have mentioned, for it to really work, you'd probably need a small fleet of drones - and more likely, data from satellite providers. In my area of work (surveillance / intelligence), we do a lot of similar stuff - students like these are awesome to work with for summer internships / thesis projects.
- yonaguska 4y agoIt's neat, but the real solution is fixing the incentives for poachers in the first place. It's an economic issue at its root