The Program for the Hatutus Fall Seminar 2011

Organized by the Pattern Recognition Society of Finland

December 1, 2011

Aalto University Otaniemi Campus

Computer Science buildingTUAS building
Konemiehentie 2, FI-02150 Espoo    Otaniementie 17, FI-02150 Espoo
Hall T5 & SaunaHall 1023

Registration: by November 18, 2011


First session, location: T5
13:00  Opening of seminar
13:05Machine vision research
Matti Pietikäinen, University of Oulu
13:35Learning Graded Relations with Kernel Methods
Tapio Pahikkala, University of Turku
14:00Coffee break
Second session, location: 1023 / TUAS
14:15Personalized voice models
Mikko Kurimo
14:40Cognitive systems research
Timo Honkela
15:05Restricted Boltzmann machines
KyungHyun Cho
15:30Multimodal video and audio analysis
Mats Sjöberg
15:55Closing of seminar
16:00General meeting, planning of SCIA 2013
location: 1023 / TUAS
17:00Dinner & sauna


Personalized voice models

With adaptive and learning methods long recording and annotation sessions are no longer required for building personalized voice models. Using this technology the recognition of everybody's speech becomes more accurate, even in noisy environment, and speech can be synthesized using anybody's voice, even in a foreign language. Personalized synthetic speech is needed when a person who has lost his voice wants to communicate using his original voice. Another important application is speech translation into a foreign language so that the speech continues to sound like the same speaker. For more information, see and