Project Topics

IUPR offers a variety of seminars and lab courses, and we supervise project courses, bachelor theses, and master theses. 

This page lists ideas and topics for smaller projects (less than six months).

For more advanced projects and research , please also see Thesis Topics

Pattern Recognition

Research in pattern recognition in our group focuses on large scale learning algorithms, parameter-free methods, and evaluation methods.

Possible projects include:
  • theoretical analysis of pattern recognition and statistical machine learning methods
  • comparative evaluations of different pattern recognition methods
  • validation and testing methods for pattern recognition systems
  • improving parameter-free learning algorithms
  • parallelizing pattern recognition algorithms with map-reduce
  • implementing high performance pattern recognition algorithms on multicore and graphics chips (CUDA, nVidia)
  • creating web-based tools for pattern recognition and image analysis
  • Seti@Home-like support for building a distributed intelligence
  • sandboxed execution of pattern recognition algorithms
  • applications of pattern recognition and machine learning to bioinformatics

Document Analysis

Our group is involved in several large document analysis projects and efforts.

Possible student projects include work in:
  • contributions to the open source OCRopus project, for example, supporting additional languages and scripts, creating new front-ends and user interface, etc.
  • capturing, enhancing, analyzing, and publishing historical documents and art
  • developing and enhancing text-in-image recognition algorithms (e.g., read street signs from a mobile or automotive camera)
  • applying machine learning to document analysis
  • creating accessiblity applications (Barrierefreiheit) through document analysis techniques

Usability and HCI

In this area, we examine a variety of usability issues in software.  Examples of small projects students can carry out are:

  • efficiency of different menu entry paradigms
  • efficiency of keyboard-based launchers (Ubiquity, Quicksilver, Gnome Do, Deskbar, Launchy, etc.)
  • understanding errors in keyboard-based launchers
  • user interfaces for code refactoring

Mobile Computing

Our research in mobile computing focuses on combining imaging (cameras) with mobile devices, data input, and location-aware services. 

Possible projects include developing mobile applications and client/server applications for
  • barcode recognition
  • location recognition from mobile cameras
  • mobile augmented reality
  • location detection based on combining multiple wireless sources and visual cues
  • an applicatinon for social location sharing
  • improving T9 (cell phone) input with language models
  • better on-screen keyboards for Android
  • pen-based input for Android
Our preferred mobile platforms are J2ME on Nokia and Android (Java).

Internet Content Analysis

Our work in internet content analysis focuses on applying content nalysis and information extraction techniques to problems that arise as part of Internet computing. 

Examples of projects are building prototypes or demonstrators for applications like
  • a mobile-accessible digital library
  • applying image retrieval techniques to automated pornography filtering
  • applying image and video databases to copyright infringement detection
  • applications of named entity recognition to address book management
  • intelligent photo albums
  • combining social network analysis and link analysis for on-line digital forensics
Projects don't have to start from scratch, but can build on a large existing set of tools for content analysis, multimedia databases and retrieval, and OCR and text analysis.