Signals and Systems (ISS) is compulsory for the 2nd year of Bc. program at FIT. The course runs in winter term and is taught by Honza Cernocky, usually in Czech on Wednesdays and in English on Fridays. It covers the bases of continuous and discrete signal processing and includes digital filtering, Fourier transforms, some 2D and random signals. See new slides, old slides, 2013 videos on Superlectures and all videos on FIT servers.
Speech signal processing (ZRE) is compulsory for "our" Master specialization "Computer graphics and Multimedia" and is taught in summer term by Honza Cernocky and other speech guys (in Czech). It covers the fundamentals (LPC, cepstrum, pitch estimation) as well as topical lectures on speech coding, recognition, speaker and language ID, etc. See slides, 2013 videos on Superlectures, and all videos on FIT servers.
Classification and recognition (IKR) is elective in any year of the Bc. program at FIT. The course runs in summer term and is taught by Lukas Burget (in Czech). It covers the bases of statistical classification and recognition - costs, errors, generative and discriminative classifiers, neural nets, SVMs, etc. See slides and all videos on FIT servers.
Advanced Classification and recognition (KRD) is elective in the PhD program. This is a seminar-style course supervised by Lukas Burget, the presentations being prepared either by him or by the students. It is run in English, usually in summer term.
Audio and Speech Processing by Humans and Machines (ASD) is elective in the PhD program. ASD is taught by Hynek Hermansky, usually as a 2-3 days intensive course in winter or summer term, in English.
We are eager to supervise good students working on speech, natural language or related (music, signal estimation, ...) Bachelor, Master and Doctoral thesis projects. We are also eager to serve as pedagogical supervisors for assignments generated by our industrial colleagues. See our PhD alumni. See all past. Bc projects and all past MSc. projects at FIT. Unfortunately, it is not possible to filter by research groups, so try to filter by the supervisor name ("Vedouci"). It needs diacritics, so you better copy the name from speech pages …
Superlectures.com is a web service and portal for lecture indexing, search and advanced multi-device presentation. The underlying technology (ASR, slide detection, etc) was developed by our group and our colleagues from the Department of Computer Graphics and Multimedia. Our spin-off company ReplayWell is responsible for the commercial operation of this service.
Prednasky.com ("prednasky" means "lectures" in Czech) serves the students of our Faculty of Information technology by providing lectures taught at FIT in indexed and accessible form (mainly in Czech). In 2015/16, FIT supported more massive replacement of the plain old video technology by Superlectures.
Talks for students
Video lecture courses
Our other speech processing video lectures
- Igor Szöke, Petr Schwarz, Honza Černocký: Dolování informací z řeči aneb "co ani Google neumí"
- Igor Szöke: Geekovo minimum umělé inteligence
- Igor Szöke: Automatické zpracování a dolování informací z audiovizuálních záznamů