Bi-Modal Emotion Recognition Dataset (RAVDESS + SAVEE + CREMA-D)
Bibliographic record
Abstract
This dataset is a visual (audio-video) resource created for bimodal emotion recognition. It is compiled from three major sources: CREMA-D (Crowd-sourced Emotional Multimodal Actors Dataset) RAVDESS (Ryerson Audio-Visual Database of Emotional Speech and Song) SAVEE (Surrey Audio-Visual Expressed Emotion) By integrating these databases, a comprehensive and diverse dataset was developed to improve the accuracy of bimodal emotion recognition systems. The resulting dataset includes audio-visual recordings of 119 actors (64 males and 55 females) representing various ethnic backgrounds and age groups. Initially, the dataset comprised 11,956 samples (6,371 from male actors and 5,585 from female actors) across seven emotion categories: Angry, Sad, Happy, Fearful, Surprised, Disgust, and Neutral. However, the dataset exhibited class imbalance, with certain emotions like "Surprised" being underrepresented. To address this imbalance, and to introduce natural variations in expression, data augmentation techniques were applied. This involved randomly selecting, duplicating, and modifying video samples from each class using methods such as pitch shifting and speed alteration. As a result, the dataset became class-balanced, containing a total of 14,257 files (7,482 from male and 6,775 from female actors). This extensive and diverse dataset improves real-world relevance by encompassing a wide range of actors, emotional expressions, ethnic diversity, and gender representation. It significantly strengthens the generalizability and robustness of emotion recognition models in practical applications. Insight of the Dataset train val test Total Total Files 10013 (5106 for Male, 4907 for Female) 2851 (1491 for Male, 1360 for Female) 1393 (880 for Male, 513 for Female) 14257 (7477 for Male, 6780 for Female) Total Actors 83 Actors (43 Male, 40 Female) 24 Actors (13 Male, 11 Female) 12 Actors (8 Male, 4 Female) 119 Actors (64 Male, 55 Female) Classwise Insights (Total Files) Training set Angry Disgust Fearful Happy Neutral Sad Surprised Male 765 705 772 777 703 768 616 Female 727 699 742 746 657 735 601 Total 1492 1404 1514 1523 1360 1503 1217 Validation set Angry Disgust Fearful Happy Neutral Sad Surprised Male 224 211 219 224 211 219 183 Female 203 190 198 203 190 198 178 Total 427 401 417 427 401 417 361 Testing set Angry Disgust Fearful Happy Neutral Sad Surprised Male 131 125 131 131 129 131 102 Female 74 67 74 74 62 74 88 Total 205 192 205 205 191 205 190 Overall Angry Disgust Fearful Happy Neutral Sad Surprised Male 1120 1041 1122 1132 1043 1118 901 Female 1004 956 1014 1023 909 1007 867 Total 2124 1997 2136 2155 1952 2125 1768
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.003 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.002 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.020 | 0.024 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".