Master Semester Project
While a large variety of filters exists to detect blob-like structures, there are much less solutions for identification of tip-like patterns. Several existing methods such as template matching were already tested but proven unsuccessful. As many microorganism in biological research are rod-shaped, such detector would be very useful in several detection and tracking application. The detector should be rotation invariant and ideally multiscale. While machine learning approaches can be tested, an filter-oriented implementation might be worth considering as such detector might be needed for fast tracking algorithms. We propose as a starting point to design, implement and experiment a detector using steerable filters.
- Virginie Uhlmann, firstname.lastname@example.org, 021 693 1136, BM 4.142
- Michael Unser, email@example.com, 021 693 51 75, BM 4.136
- Zsuzsanna Püspöki, firstname.lastname@example.org, BM 4.139, Tel: 021 693 51 57