Frank R. Schmidt


In the past, we were working on a range of topics in Computer Vision, Image Processing and Pattern Recognition. During this work, several different data were accumulated that we are happy to share with other researchers. Please refer to the respective publication when using this data.

Shape Prior Database

The archive contains a dynamic shape prior and a test sequence stored in different directories:

This directory contains a set of shapes that are temporally sorted. This order can be used in order to learn a dynamical shape prior.
This directory contains the frames of a video stored as images. The images are ordered temporally.
Shape Priors in Variational Image Segmentation: Convexity, Lipschitz Continuity and Globally Optimal Solutions


Planar Graph Cut Database

The archives and contain one picture at different resolutions (from 0.05 to 4.00 Megapixels). Each image pair defines a planar graph as follows :

The image is at a resolution of approximately X.XX Megapixels. The height and width of an image defines a rectangular 4-connected planar grid. In our work, we used a capacity of exp(-(I-J)2) for every edge, where I and J describes the color value of neighboring pixels.
This image encodes the preselected foreground (red) and background (blue) that are associated with the source and sink resp.
Efficient Planar Graph Cuts with Applications in Computer Vision