M2ASR-MONGO: A Free Mongolian Speech Database and Accompanied Baselines

This paper describes a logatome discrimination test for the assessment of speech perception in cochlear implant users (CI users), based on a multilingual speech database, the Oldenburg Logatome Corpus, which was originally recorded for the comparison of human and automated speech recognition. The logatome discrimination task is based on the presentation of 100 logatome pairs (i.e., nonsense syllables) with balanced representations of alternating “vowel-replacement” and “consonant-replacement” paradigms in order to assess phoneme confusions. Thirteen adult normal hearing listeners and eight adult CI users, including both good and poor performers, were included in the study and completed the test after their speech intelligibility abilities were evaluated with an established sentence test in noise. Furthermore, the discrimination abilities were measured electrophysiologically by recording the mismatch negativity (MMN) as a component of auditory event-related potentials. The results show a clear MMN response only for normal hearing listeners and CI users with good performance, correlating with their logatome discrimination abilities. Higher discrimination scores for vowel-replacement paradigms than for the consonant-replacement paradigms were found. We conclude that the logatome discrimination test is well suited to monitor the speech perception skills of CI users. Due to the large number of available spoken logatome items, the Oldenburg Logatome Corpus appears to provide a useful and powerful basis for further development of speech perception tests for CI users.

Download Full-text

Multi-Path and Group-Loss-Based Network for Speech Emotion Recognition in Multi-Domain Datasets

Sensors ◽

10.3390/s21051579 ◽

2021 ◽

Vol 21 (5) ◽

pp. 1579 ◽

Cited By ~ 1

Author(s):

Kyoung Ju Noh ◽

Chi Yoon Jeong ◽

Jiyoun Lim ◽

Seungeun Chung ◽

Gague Kim ◽

...

Keyword(s):

Emotion Recognition ◽

Short Term Memory ◽

Domain Adaptation ◽

Classification Model ◽

Speech Emotion Recognition ◽

Target Domain ◽

Model Generalization ◽

Speech Database ◽

Emotion Labels ◽

Temporal Feature

Speech emotion recognition (SER) is a natural method of recognizing individual emotions in everyday life. To distribute SER models to real-world applications, some key challenges must be overcome, such as the lack of datasets tagged with emotion labels and the weak generalization of the SER model for an unseen target domain. This study proposes a multi-path and group-loss-based network (MPGLN) for SER to support multi-domain adaptation. The proposed model includes a bidirectional long short-term memory-based temporal feature generator and a transferred feature extractor from the pre-trained VGG-like audio classification model (VGGish), and it learns simultaneously based on multiple losses according to the association of emotion labels in the discrete and dimensional models. For the evaluation of the MPGLN SER as applied to multi-cultural domain datasets, the Korean Emotional Speech Database (KESD), including KESDy18 and KESDy19, is constructed, and the English-speaking Interactive Emotional Dyadic Motion Capture database (IEMOCAP) is used. The evaluation of multi-domain adaptation and domain generalization showed 3.7% and 3.5% improvements, respectively, of the F1 score when comparing the performance of MPGLN SER with a baseline SER model that uses a temporal feature generator. We show that the MPGLN SER efficiently supports multi-domain adaptation and reinforces model generalization.

Download Full-text