Kassel State of Fluency - Challenge

The stuttering Kassel State of Fluency corpus KSoF-C, provided is derived from<br> the Kassel State of Fluency (KSoF) corpus. The original corpus fea-<br> tures some 5500 typical and nontypical (stuttering) 3-sec segments<br> from 37 German speakers with an overall duration of 4,6 hours. The<br> segments contain speech of persons who stutter (PWO). The record-<br> ings from which the segments were extracted were recorded before,<br> during, and after PWOs underwent stuttering therapy. KSOF-C only features the 4601 non-ambiguously labeled segments.<br> The task proposed in this challenge is the classification of speech<br> segments as one of the 8 classes - the seven stuttering-related classes<br> mentioned above and an eighth “garbage” class, denoting unintelligible segments, segments containing no speech, or segments that<br> are negatively affected by loud background noise. The dataset is split by speaker (train, 23 spk, devel, 6 spk, test, 8 spk)