Building Computer Vision Systems with Martin Jul
While at NDC is Oslo, Carl and Richard talk to Martin Jul about the state of visual computing technology today. As Martin says, the future is very bright - amazing free libraries are now available that make it simple to do image and character recognition with all sorts of cameras, even one on the Raspberry PI! The conversation digs into a project that Martin built to identify when the pizza guy is at his front door, exploring what it takes to recognize a pizza box in video. There are digital cameras everywhere today, shouldn't we be building apps to take advantage of them?
Guests:
Martin Jul
Martin Jul is a software craftsman working in startup stealth mode in the field of computer vision and machine learning. He has spent his career building applications for web, mobile and data centres in fields as diverse as dating, defense, finance and telecom. Before the arrival of the Internet Age he studied mathematics and computer science in Copenhagen, and he is very happy that libraries and powerful processors are finally making the hitherto exotic academic disciplines for computing with images accessible to everyone. He lives in Copenhagen, Denmark.
Links:
- Vivino on iTunes https://itunes.apple.com/gb/app/vivino-wine-scanner/id414461255?mt=8
- Microsoft Research Face SDK http://research.microsoft.com/en-us/projects/facesdk/
- OpenCV http://opencv.org/
- OpenCV for .NET http://www.nuget.org/packages/OpenCV.Net/3.2.2
- Raspberry PI http://www.raspberrypi.org/
- Martin's Blog http://startupmode.dk/
- Tesseract https://code.google.com/p/tesseract-ocr/
- emgu library http://www.emgu.com/wiki/index.php/Main_Page
- GaussianFace http://arxiv.org/abs/1404.3840
- Deep Face Recognition http://www.faceplusplus.com/wp-content/uploads/2014/03/Learning-Deep-Face-Representation.pdf