TOPICS FOR IT-PROJECT + MSc-THESIS IN THE FIELD OF COLOR SCIENCE (time period is May/June - 31.12.2007) In some of the topics there will be a 700 euros/month scholarship. These topics will be decided later, when InFotonics Center will get decisions about the financing. You may apply now according to your preferences. 1. Title: Image colorization using spectral colors Supervisor: Vladimir Botchko Special: Nonlinear dimensionality reduction technique including Locally Linear Embeddings (LLE) and ISOMAP will be tested for purposes of image colorization. 2. Title: Spectral image segmentation using Markov Random Walk Model Supervisor: Vladimir Botchko Special: Novel algorithm will be tested. Most probable state sequence of features will be defined using Markov Random Walk on data. Then for this sequence initial, transition and emission probabilities will be computed. Then these model parameters will be used in pixel classification (Viterbi algorithm). 3. Title: Application for museums: color constancy in image retrieval Supervisor: Vladimir Botchko Special: Color depends on illumination. This makes it difficult to use in applications. Color constancy algorithm with a chromagenic filter will be tested. Application is either with mobile phones or digital camera. 4. Title: Dimensionality estimation in specral refletances and spectral images Supervisor: Vladimir Botchko Special: Projections based on nonlinear transformation criteria (ISOMAP) and Spectral Fractal Dimension will be tested to define a dimensionality of spectral data in data sets and images. 5. Title: The measurement accuracy in the 3D spectral imaging Supervisor: Birgitta Martinkauppi Special: In the color lab there is a new laser-based 3D image retrieval system. One MSc thesis has been done where spectral image was rendered over 3D shape. Now the system's imaging capabilities will be studied further. 6. Title: Characterization and tuning of a printer Supervisor: Birgitta Martinkauppi Special: Accurate and visually good quality printing is an important aspect of a printer. Spectral color methods give a good tool for managing printers. Here, methods for this will be developed. 7. Title: Contrast sensitivity of eye and color diplays Supervisor: Hannu Laamanen Special: The topic is related to the management system development for different type of display systems 8. Title: Computational determination of spectral sensitivity curves of different imaging systems Supervisor: Ville Heikkinen Special: This is quite mathematical topic which should lead application tests in some of the color group practical projects