Emotion Recognition: Using Emotion Datasets to Enhance your AI Performance

Speech emotion recognition generally refers to the process by which a machine automatically recognizes human emotions and emotion-related states from speech.

The challenge of speech emotion recognition

● Data labeling is a very time-consuming and labor-intensive process that requires a large number of professionals. Although there are many emotion datasets available, it is very difficult for researchers to build emotion datasets for specific scenarios.

● Feature extraction and selection is still a hard problem. Due to the variety of speakers, changing emotions, and different lengths of speech clips, the artificially selected features cannot cover all the information, and the robustness to the overall data is not strong enough.

● The field of speech emotion recognition is still relatively young and lacks official standards. Different people have different opinions on the same speech emotion. At the same time, a piece of speech often contains multiple emotions and is highly subjective, which leads to the fact that the results of many current studies are not universal.

Facing the problem of scarcity of emotion recognition training corpus, Nexdata has developed a series of emotion datasets.

20 People-English Emotion Dataset by Microphone

English emotion dataset captured by microphone, 20 American native speakers participate in the recording, 2,100 sentences per person; the recorded script covers 10 emotions such as anger, happiness, sadness; the voice is recorded by high-fidelity microphone therefore has high quality; it is used for analytical detection of emotional speech.

1,003 People – Emotion Dataset

1,003 People - Emotional Dataset. The emotion dataset diversity includes multiple races, multiple indoor scenes, multiple age groups, multiple languages, multiple emotions (11 types of facial emotions, 15 types of inner emotions). For each sentence in each video, emotion types (including facial emotions and inner emotions), start & end time, and text transcription were annotated. This emotion dataset can be used for tasks such as emotion recognition and sentiment analysis.

In addition to these finished emotion dataset, Nexdata also provides customized data collection and labeling services to help customers overcome data difficulties in the field of sentiment analysis.

End

If you want to know more details about the emotion datasets or how to acquire, please feel free to contact us: [email protected].