Methods for precise local affine frame constructions on MSERs

• Thesis title in Czech: Metody pro přesnou konstrukci lokálních afinních rámců na MSER
• Thesis title in English: Methods for precise local affine frame constructions on MSERs
• Key words: MSER, LAF, Lokálne afinní rámce, feature detektor
• English key words: MSER, LAF, Local affine frame, feature detector, contour refinement
• Academic year of topic announcement: 2013/2014
• Thesis type: rigorosum thesis
• Thesis language: angličtina
• Department: Department of Software Engineering (32-KSI)
• Author: hidden - assigned and confirmed by the Study Dept.
• Date of registration: 16.09.2014
• Date of assignment: 16.09.2014
• Confirmed by Study dept. on: 16.09.2014
• Date and time of defence: 01.12.2014 00:00
• Date of electronic submission: 17.09.2014
• Date of submission of printed version: 16.09.2014
• Date of proceeded defence: 01.12.2014

Guidelines

1. Familiarize yourself with the detector of distinguished regions MSER (Jiri Matas, Ondrej Chum, Martin Urban, Tomás Pajdla: Robust wide-baseline stereo from maximally stable extremal regions. Image Vision Comput. 22(10): 761-767 (2004)
2. Familiarize yourself with evaluation methodology for MSER precision and repeatability on standard datasets.
(Krystian Mikolajczyk, Tinne Tuytelaars, Cordelia Schmid, Andrew Zisserman, Jiri Matas, Frederik Schaffalitzky, Timor Kadir, Luc J. Van Gool: A Comparison of Affine Region Detectors. International Journal of Computer Vision 65(1-2): 43-72 (2005)
3. Familiarize yourself with the concept of Local Affine Frames (Stepán Obdrzálek, Jiri Matas: Object Recognition using Local Affine Frames on Distinguished Regions. BMVC 2002)
4. Propose possible improvements for Local Affine Frame construction on MSER
5. Implement and evaluate the proposed algorithms.

References

1. Hartley, Zisserman: "Multiple View Geometry in Computer Vision";
2. Ponce, Forsyth: "Computer Vision"
3. The papers cited above