Hakan Bilen
Hakan Bilen earned degrees at Sabanci and a doctorate at KU Leuven, then conducted postdoctoral work with Andrea Vedaldi at Oxford before joining Edinburgh. His doctorate studied image classification with latent parameters, including weakly supervised object detection and action recognition. At CVPR 2016, Bilen and Vedaldi presented Weakly Supervised Deep Detection Networks, an architecture that selects and classifies image regions while training from image labels. He leads Edinburgh’s Visual Computing Group and has taught Machine Learning Practical since 2018.
- Now at
- University of Edinburgh
- Works
- 116
- Citations
- 4,706
- h-index
- 28
Publication numbers from OpenAlex, as of 14 October 2026.
Timeline
-
Earlier career
From robotics to visual learning
Bilen earned bachelor’s and master’s degrees at Sabanci and a doctorate at KU Leuven. The doctorate studied image classification with latent parameters, weakly supervised detection, and action recognition.
-
Earlier career and 2016
Oxford and detection networks
After his doctorate, Bilen worked with Andrea Vedaldi in Oxford’s Visual Geometry Group. Their paper Weakly Supervised Deep Detection Networks appeared at CVPR in 2016. It addressed object detection using deep convolutional networks with image-level supervision.
-
2016
Learning locations from image labels
The paper’s architecture selects and classifies image regions while being trained as an image classifier. The authors reported comparisons on PASCAL VOC. The method learns object locations from limited annotation.
-
2018 and later
Visual computing at Edinburgh
Bilen has taught Machine Learning Practical since 2018. He is a reader in Edinburgh’s School of Informatics and leads the Visual Computing Group.
Sources and credits
Photo: homepages.inf.ed.ac.uk; creator not stated. License: Not stated by source. Source: homepages.inf.ed.ac.uk. Cropped to a square and resized. Ask us to remove or replace this photo.