As always, I am more upbeat in describing new technologies. But in the end, it is your own project to do the real judgment. Applying object detection in home devices is getting traction. So this technology deserves some attention.

Just one bad comment on most AI devices. Companies should hire better industrial product designers and stop making products look like an air freshener or video camera in the 90s. lol.

For productization, my fundamental question is whether those projects have created a “must have” story rather than just a few sound good use cases from the PM.

Above is just a general comment on AI devices rather on DeepLens. So this is more specific to your question. https://github.com/apache/incubator-mxnet/tree/master/example will give you some ideas of what algorithms may available.

For classification, you can train it with your own data and upload the model to the camera. In object detection, it focuses on SSD with ResNet or VGG for now. But this is highly evolving, so this will change quickly and fast. I kind of understand why they pick SSD first and make sense to me. When they pick SSD, it shows the intention of using it for real-time detection. But the device has a regular CPU only, it may be good for inference but is different from what Apple approaches AI with its Bionic chip. Other companies also make pre-trained models available for developers. The solution approaches, execution details and platform are quite different.

Written by

Deep Learning

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store