Viva Las Vegas!

Sightseeing on the Las Vegas Strip with our demo system

We’re in Las Vegas right now, exhibiting at CES 2023! It’s been a great experience so far and we’re looking forward to the rest of the show!

As you can imagine, the whole Ubicept team has been incredibly busy talking to prospective partners and customers. Still, we decided to take a break for a sightseeing drive down the Strip since some of us haven't seen the city before. Of course, we brought one of our demo systems along for the ride!

Here’s a side-by-side comparison with a high-end automotive camera:

As you can see, the Ubicept solution can capture significantly more detail. We see all the people enjoying the scenery—and so can an off-the-shelf object detection algorithm! Here’s a comparison:

Strangely, the blurry input from the automotive camera results in the object detector seeing multiple cars and no people! Here’s a closeup of the “cars” in question:

Our human perception systems can actually infer the shapes of people here based on context, so you might wonder whether a better object detection algorithm could as well. It’s certainly possible, but key takeaway here is that the Ubicept solution makes it unnecessary for humans (or computers) to strain their eyes (sensors) or brains (processors) to perform perception tasks, which significantly reduces the likelihood of errors. This is especially useful in more challenging situations like this one:

With the automotive camera, it’s extremely difficult to see the people waiting at the bus stop due to the motion blur and wide dynamic range of the scene. But with the Ubicept solution, we can even see what the people are doing! And here’s the object detector again:

Here are the full video captures:

We’ll be following up with more demos later. In the meantime, if you’d like to know more about how Ubicept can help solve your most challenging perception problems, reach out or drop by our booth if you also happen to be at CES!

Previous
Previous

Ubicept at CES 2023

Next
Next

Perception with Extreme Dynamic Range