Computer Vision News Computer Vision News 44 Congrats, Doctor Denys! Denys Rozumnyi received his PhD at ETH Zurich under the supervision of Marc Pollefeys. He is also a Research Fellow at the Czech Technical University in Prague, working with Jiri Matas. During his PhD, he interned at Google Research and Meta Reality Labs. His master thesis received the best master thesis award in the Czech Republic. Now, he is an incoming Research Scientist at Meta Reality Labs. Motion blur is a common cause of degraded image quality in video and photography. The primary sources of image blur are rapid camera motion, fast object motion, long exposure times due to low light settings, and a combination of these effects. In such cases, classical methods, and even humans, often fail to recover the object’s appearance and motion. In his thesis, Denys proposed a comprehensive range of methods to address the challenges associated with motion-blurred objects, i.e. retrieval, detection, tracking, trajectory estimation, deblurring, 3D reconstruction, and the intersection of several of these tasks at once. First, we are interested in locating frames where a motion-blurred object occurs (retrieval). When such frames are located, the next question is where exactly is this object in this frame (detection). Next, we are interested in where this object is in the following frames or where it is going (tracking). Then, imagine that this object must be identified, but as mentioned above, the object is severely blurred due to its own motion. Thus, the next task is to deblur an image of this object (deblurring) to know how the object would look like if it would have been recorded by a high-speed camera. For a more precise
RkJQdWJsaXNoZXIy NTc3NzU=