TensorCam is here to change our life! In this article I’ll describe the most powerful easy-to-use deep learning enabled camera I’ve seen so far. And also it’s open for developers, with this camera people of all skill levels can get started with deep learning in less than 5 minutes.

It is time for a new development in the world of computer vision and Deep Learning, and this is what the folks of Deep Cognition are launching for us.
If you want to practice your Deep Learning skills this is an amazing place to start, with this camera you can build your own computer vision applications in minutes.
Here’s a great intro video of the system:
The best part is that it fully connects with Deep Learning Studio (DLS),where you can choose from a simple but powerful GUI where you can drag and drop neural networks and create Deep Learning models with AutoML, to a full autonomous IDE where you can code and interact with your favorite libraries.
DLS simplifies and accelerates the process of working with deep learning across popular frameworks such as TensorFlow and MXNet. You will be able to push your deep learning models from Deep Learning Studio to TensorCam. Moreover, TensorCam will come with full software SDK.
This would be your workflow:

If you want to know more about DLS check this out:
and this great article by Rajat:
Deep Learning made easy with Deep Learning Studio – Complete Guide
Features:
- Built-in NLP technology, which will allow you to develop custom language based notifications. These notifications will be delivered through smart home devices like Alexa.
- Built-in state of the art deep learning chip, which will allow you to run deep learning models at a whopping 2.4 Terra-OPS computing performance.
- Its tight integration with Deep Learning Studio (DLS) makes it highly flexible and customizable, which will allow you to run your own deep learning models in a matter of hours.

Here you can see how it works with Alexa:

You can create a model for shelf inventory monitoring:

A model that can start recording when your baby moves (yeah you can do that right now):

And many more 🙂
Comparison to other cameras
Their deep learning chip with 2.4 Tera Operations Per Second (TOPS) it’s also the highest of the market right now as you can see in this comparison:

And don’t get me wrong, DeepLens and the Google Vision Kit are awesome too, I actually have used the Google Vision Kit in the past, and the only part I wish it was easier back them was the Programming and interface. I mean you have to have everything in a RaspberryPi to then create a model, train it and hope for the best that it will run there. Also although the camera was very good it didn’t move.
And that’s a big difference with TensorCam because it comes with PTZ (pan–tilt–zoom) meaning that the camera is capable of remote directional and zoom control. Which is awesome and necessary for so many applications.
Here’s a more in-depth comparison:

What can you build with TensorCam?

You can get started by using the TensorCam sample projects, which cover some of the most popular computer vision use cases. As your skills and ideas develop, you can build custom deep learning models using Deep Learning Studio’s cloud or desktop version. As a developer, here are some of the examples that you would be able to build and deploy on TensorCam with minimal or no-coding:

If you want more ideas checkout these ones here:
Here you can see a full video example of an Employee Tracking system (so you know who really shows up):
Oh and you also have night vision if you are wondering:
The KickStarter Campaign

Now for the most important part, if you are amaze (you should), these guys needs our help to fund their project. This is the first time I see something like this in a KickStarter campaign so please make sure to enter the webpage and support the project:
https://www.kickstarter.com/projects/deepconition/tensorcam-worlds-first-ai-camera-with-ptz-for-deve
If you are unable to do it please tell your friends and company, you can get a big discount if you are able to make a pledge right now. A final words from the creators:
We anticipate the on-time delivery for TensorCam in March 2019. We’re wrapping up the product development phase for the main features, but we’d still like to hear great ideas to implement from the Kickstarter backers, about some extra features which we can include (as long as it doesn’t impact the delivery date). We’ve found reliable manufacturing partners for the camera and AI-board, but even with the best manufacturing processes setup, unexpected hiccups can happen. Logistics issues are also being solved to ensure smooth delivery for when we’ve exceeded our goal.
Thanks for reading this. I hope you found something interesting here 🙂 If you have questions just follow me on Twitter:
and LinkedIn:
See you there 🙂