Deep Learning in Computer Vision

Explore deep learning in computer vision through this 48-minute conference talk by Krzysztof Kudrynski and Błażej Kubiak from TomTom. Delve into the fascinating world of machine image understanding, starting with basic machine learning concepts and progressing to advanced practical tips and state-of-the-art architectures. Discover how complex problems can be solved using portable computers and gain insights into the reasons behind adding new layers and introducing new concepts. Learn about designing and developing neural networks that can interpret visual information, with examples ranging from edge detection to self-driving cars. Gain expert knowledge on artificial intelligence, data analysis, computer vision, and robotics from speakers with extensive experience in software design and big data processing. Follow the journey from science fiction to reality as the talk covers topics such as convolutional networks, backpropagation, specialized networks, and the Chaos framework, concluding with a demo and practical applications in the field.

Introduction
Agenda
Science Fiction
Around the corner
EagleEye
Microsoft Autobahn
David Hasselhoff
Fiction
Selfdriving cars
Edge detection
Tensor flow
Implementation
Accuracy Cost
Mysterious Filter
Results
Tensorflow
Dot
Final result
Final results
Summary
Artificial Intelligence
convolutional network
backpropagation
the distant future
the Jaeger
classification problem
goal
specialized networks
training networks
Chaos framework
Chaos initializing
Stencilboard
Demo
Short introduction
Upper application
Conclusion