Advancements in robotic navigation, mapping, object search and recognition rest
to a large extent on robust, efficient and scalable semantic understanding of
the surrounding environment. In recent years we have developed several approach
es for capturing geometry and semantics of environment from video, RGB-D data,
or just simply a single RGB image, focusing on indoors and outdoors
relevant for robotics applications.
I will demonstrate our work on detailed semantic parsing and 3D structure
using deep convolutional neural networks (CNNs) and object detection and object
pose recovery from single RGB image. The applicability of the presented
for autonomous driving, service robotics, mapping and augmented reality
applications will be discussed.