Visual tracking plays an important role in many computer vision tasks. A common assumption in previous methods is that the video frames are blur free. In reality, motion blurs are...
Thispaper presents a text word extraction algorithm that takes a set of bounding boxes of glyphs and their associated text lines of a given document andpartitions the glyphs into ...
We introduce the challenge problem for generic video indexing to gain insight in intermediate steps that affect performance of multimedia analysis methods, while at the same time...
Cees Snoek, Marcel Worring, Jan van Gemert, Jan-Ma...
Most multi-camera systems assume a well structured environment to detect and track objects across cameras. Cameras need to be fixed and calibrated, or only objects within a traini...
Alexandre Alahi, Pierre Vandergheynst, Michel Bier...
Abstract. This paper introduces a new model-based approach for simultaneously reconstructing 3D human motion and full-body skeletal size from a small set of 2D image features track...