Bachelor Informatica Afstudeer Symposium

Op 1 juli is het Bachelor Informatica Afstudeer Symposium (BIAS). Vanaf 12:45 is eenieder welkom in zaal 253 van de Bernoulliborg.

Programma Bachelor Informatica Afstudeer Symposium (BIAS) 2011, 1 juli 2011

Zaal: Bernoulliborg 253

12:45 - 13:00 Ontvangst in zaal

13:00 - 13:05
Opening -- Michael Wilkinson

13:05 - 13:30

In-line Editing: a New Approach to Editing Wikis

Jan Paul Posma

Abstract: We propose some new interfaces for editing wikis. In these interfaces the original page is shown, with editable elements in the page highlighted, such as sentences, sections, media, references, and templates. It is then possible to edit the original wikitext by clicking these elements. Such an interface is generated by the existing parser, which solves the problem of having to port often complex parsers — without a formal grammar — to client side code. This is currently a large problem for some wikis, including Wikipedia. We discuss three different interfaces and usability testing with them. We also look at how to leverage an existing parser to generate these interfaces.

13:30 - 13:55

Morphological attribute filtering and visualization with 3-D salience trees

Roan Kattouw

The salience tree is a data structure representing an image at multiple levels of detail. I was originally designed as a faster alternative to the binary partition tree, and used for object recognition in 2-D images. I have extended the salience tree to operate on 3-D volumes, and used it for morphological attribute filtering and visualization. Comparing the salience tree to the max-tree, a data structure often used in this field, it appears that while the salience tree does not suffer from some of the max-tree's theoretical defects, it does not perform adequately in realistic cases where these defects are not a factor.

13:55 - 14:20

Moment Based Text Recognition

Steven Bouma, Ruurd Moelker, Werner Buck

Traditional character recognition (OCR) uses thresholding and pre-defined characters to recognize text. Our study theorises that thresholding loses potentially valuable information. On that basis our study attempts to provide a more versatile approach to character recognition using a max-tree and image moments with minimal thresholding. The max-tree is a tree representation of the image where where each level represents a grey-level and each node contains attributes. The image moments are scale and translational invariant, this allows recognition of a wide range of characters instances. Finally supervised learning is used to classify the characters, and line and word segmentation is used to obtain a full textual reconstruction of an input image.

14:20 - 14:45

Boomringherkenning

Boombiologen en archeologen werken wel eens met restanten van boomstammen die ze in de grond vinden bij opgravingen. Informatie over de breedte van de boomringen in deze stam zijn daarvoor belangrijk. Daarom probeer ik de ringen in een boomstam vinden door een afbeelding van een zo'n stamdoorsnede te filteren op ringen. Uiteindelijk moet dit geautomatiseerd informatie over de ringen verscha en en in een database met eventueel meer door archeologen en biologen ingevoerde informatie opslaan.

14:45 - 15:00

Pauze

15:15 - 15:30

Uitreiking Best Speaker Award BIAS 2011