AffectNet is the largest database of facial expressions, valence, and arousal in the wild, designed for automated facial expression recognition in both categorical and dimensional emotion models. It contains over 1M facial images collected from the internet using 1250 emotion-related keywords in six languages. Around 440K images were manually annotated for seven discrete facial expressions and the intensity of valence and arousal. Baseline deep neural networks were trained to classify expressions and predict valence/arousal, outperforming conventional machine learning methods.
View Database
DISFA (Denver Intensity of Spontaneous Facial Action Database) is a non-posed facial expression database designed for automatic action unit detection. It contains stereo videos of 27 adult subjects (12 females and 15 males) with different ethnicities, captured using a PtGrey stereo imaging system at high resolution (1024×768). The intensity of AU’s (0-5 scale) was manually scored by two human FACS experts, and the dataset includes 66 facial landmark points per image. The database is available for research purposes.
View Database
DISFA+ (Extended Denver Intensity of Spontaneous Facial Action Database) is an extension of DISFA with both posed and non-posed facial expression data for the same individuals. It includes manually labeled frame-based annotations of 5-level intensity for twelve FACS facial actions and provides facial landmark points along with self-reports on posed expressions. This dataset is ideal for studying differences between posed and non-posed action units and facial expression dynamics.
View Database
RyanSpeech is a high-quality male speech corpus designed for text-to-speech (TTS) research. Unlike many publicly available TTS corpora, it features over 10 hours of a professional male voice actor's speech recorded at 44.1 kHz, derived from real-world conversational settings. Optimized for developing TTS systems in real-world applications, RyanSpeech has been used to train state-of-the-art speech models, achieving a 3.36 mean opinion score (MOS) in its best model.
View Database
Fingerprint-based positioning is emerging as a reliable alternative to GPS and cellular localization in urban areas. However, the lack of public datasets limits research in this field. To address this, we introduce OutFin—a comprehensive outdoor fingerprint dataset collected using two smartphones. It includes WiFi, Bluetooth, and cellular signals, along with sensor data from the magnetometer, accelerometer, gyroscope, barometer, and ambient light sensor. Data was gathered from 122 reference points across four distinct sites, each with unique GNSS visibility and layout. Prior to release, OutFin underwent rigorous testing to ensure data quality, making it a valuable resource for advancing fingerprint-based positioning research.
View Database
