Hello nerdz,
Being a fan of TED Talks, I found a really interesting one about AI and object recognition. Joseph Redmon is presenting in this show his work about the Darknet. It’s an open source C code which is an AI that can be trained to recognize any object, and it’s damn fast!
Here is a link to the TED talk about Darknet: Joseph Redmon TED talk
If you want a quick introduction, I recommend you read his webpage: YOLO Darknet
This is the fastest way to test it under Linux:
git clone https://github.com/pjreddie/darknet cd darknet make
This will compile the code on your machine. It can take a while depending on the device (my first test was on my Android).
The next step is to download the pre-trained “brain”:
wget https://pjreddie.com/media/files/yolo.weights
And now you can run a test:
./darknet detect cfg/yolo.cfg yolo.weights data/dog.jpg
The result should looks like this:
You can now follow his blog post and run the Darknet live with a webcam like in the TED presentation.
I think you will agree it’s pretty impressive and easy to configure without a complex setup.
Do you want to know more about AI’s, look at my post about TensorFlow.
