Speech sounds (=acoustics) are produced as the coordinated movement of the speaking organs (=articulation). There are several available methods to model the relation of articulatory movements and the resulting speech. The reverse problem is often called as acoustic-articulatory inversion. The task of the student is to investigate available algorithms and develop new methods based on 2D ultrasound using modern machine learning methods (e.g. deep neural networks). Suggested programming languages: C / C++ / C# / Python. The research will be conducted in collaboration with the MTA-ELTE Lingual Articulation Research Group (Momemtum grant, http://lingart.elte.hu) and with the Industrial University of Santander (Bucaramanga, Columbia).

Speech sounds (=acoustics) are produced as the coordinated movement of the speaking organs (=articulation). There are several available methods to model the relation of articulatory movements and the resulting speech. The reverse problem is often called as acoustic-articulatory inversion. The task of the student is to investigate available algorithms and develop new methods based on 2D ultrasound using modern machine learning methods (e.g. deep neural networks). Suggested programming languages: C / C++ / C# / Python. The research will be conducted in collaboration with the MTA-ELTE Lingual Articulation Research Group (Momemtum grant, http://lingart.elte.hu) and with the Industrial University of Santander (Bucaramanga, Columbia).