Speech Recognition Systems

This course deals with the efficient coding of still image and video sequence and the international standards for transmission and storage of image information. Topics cover the representation of image signals, sampling, quantization, entropy coding, predictive coding, transform coding, subband coding, vector quantization, motion estimation, motion-compensated coding, segmentation-based coding, various international standards for bi-level image coding, still image coding and video coding.

This course provides basic theory and techniques for the representation and processing of digital video. Topics include digital video formats, video spatiotemporal Sampling, 2-D/3-D motion estimation, motion segmentation, digital video filtering, video enhancement, video compression, and digital video system. In addition to the theory, students suppose to participate in experiments that are related to the above topics.

This course offers the basic mathematical backgrounds and implementation techniques of not only recent mobile speech coding methods including CELP but also audio coding techniques such as MP3 and AAC. In addition, we study the trends for convergence of speech and audio coding techniques. (Prerequisite: EE432)

This course deals with fundamental concepts of multiple view geometry for 3D computer vision, such as projective geometry, transformation, estimation of the transformation parameters, camera model and camera matrix, epipolar geometry, fundamental matrix, trifocal tensor, and 3D Structure computation, and so on.