I took part in TSA/Kaggle's competition a year ago. https://www.kaggle.com/c/passenger-screening-algorithm-challenge/leaderboard https://www.kaggle.com/c/passe

I took part in TSA/Kaggle's competition a year ago. https://www.kaggle.com/c/passenger-screening-algorithm-challenge/leaderboard https://www.kaggle.com/c/passenger-screening-algorithm-chall... Let me first say that how well these things work is not a property of the scanners alone, but also of the algorithms they run. I'm quite certain that they can be made to work much better than a random choice, unlike what some other posters claim. The competition did have some issues. For example, the subjects were re-used in multiple scans (although not between training and test2 datasets), and the algorithms were allowed to exploit that (perhaps an oversight in the setup), but most entrants did not actually exploit this fact. Also the dataset was quite small (~20 subjects). We were still doing much better than random. With a larger dataset and perhaps better scanner design, this would work even better.