What’s this about computer vision and sensor fusion? They aren’t just putting cheapo RFID tags on everything and detecting them on the walk out? (maybe it’s both, I guess)
Makes me think there is room for error on that bill on the way out. Can you imagine arguing with someone in customer service about “No, I didn’t end up taking that cupcake, I put it back on the shelf. Look, I know you have video of me taking it, I’m saying you’re missing the video where I put it back. No, I didn’t misplace it.” I wonder if you have to put everything back right where you found it. I wonder if there are people to review the algorithm’s video->customer action decisions as they happen to guide the algorithm.
Yes, I realize there is room for error with people involved as well. Maybe even a greater chance for error. Also, my background Master’s and ABD research in grad school was computer vision, this is all fascinating to me.