Generating dance motion using musical features
Bibliographic record
Abstract
In a computer animation workflow, an animator has to plot keyframes and adjust the in-between frames to create or edit a motion, and motion is represented by keyframes and in-between frames that connect the keyframes.The sequence of significant poses, represented by keyframes, determine the general motion in animations.In-between frames are generated by interpolation strategies using keyframes, such as linear interpolation or parametric curve equation, where the starting and ending point of the parametric curve set to the two adjacent selected keyframes.In-between frames determine the trajectory and the speed an animated object moves from one keyframe to the next.Since in-between frames are reconstructed from adjacent keyframes, the selection of keyframes influences the reconstruction distance of reconstructed in-betweens, and a great amount of research has focused on strategies to select appropriate keyframes.Dance motion and dance music are often closely related to each other.Choreography is generally designed to synchronize with the rhythm of the music, and various deep learning research topics have been proposed to choreograph dance motion from dance music.In this thesis, inspired by previous research studying the close relation between dance and music, I use musical features of dance music to select keyframes and examine the reconstruction distance of Abstract ii the generated dance motion based on the AIST++ dataset.I hypothesize that the close relation between dance and music may improve reconstruction of dance motions.Three experiments are designed to evaluate the effect of musical features when reconstructing dance motions.The first experiment is used as a baseline and does not use musical features in both keyframe selection and in-between reconstructions.Keyframes are selected evenly over the measure.The List of Tables xiii 4.2 Evaluate F ID k with three, five, and nine keyframes. . . . . . . . . . . . . . . .4.3 Evaluate M SE with three, five, and nine keyframes. . . . . . . . . . . . . . . . .4.4 Evaluated results of dance reconstructed by selecting nine keyframes and reconstruct in-between frames without musical features. . . . . . . . . . . . . . .4.5 Evaluated results of dance reconstructed by selecting three keyframes and reconstruct in-between frames without musical features. . . . . . . . . . . . . . .4.6 Results of selecting nine keyframes with keyframe selection strategies using musical features, and reconstruct in-between frames using strategies independent of musical features. . . . . . . . . . . . . . . . . . . . . . . . . . . .4.7 Results of selecting three keyframes with keyframe selection strategies using musical features, and reconstruct in-between frames using strategies independent of musical features. . . . . . . . . . . . . . . . . . . . . . . . . . . .4.8 The evaluated result of dances reconstructed by evenly selecting nine keyframes and deriving in-between frames using musical features. . . . . . . . . . . . . . .4.9 The evaluated result of dances reconstructed by evenly selecting three keyframes and deriving in-between frames using musical features. . . . . . . . . . . . . . .4.10 Evaluated results of dance reconstructed by selecting nine keyframes and reconstruct in-between frames with/without musical features. . . . . . . . . . . .4.11 Evaluated results of dance reconstructed by selecting three keyframes and reconstruct in-between frames with/without musical features. . . . . . . . . . . .List of Tables xiv 4.12 Evaluated results of dance reconstructed by selecting nine keyframes and reconstruct in-between frames with/without musical features. . . . . . . . . . . .83 4.13 Evaluated results of dance reconstructed by selecting three keyframes and reconstruct in-between frames with/without musical features. . . . . . . . . . . .
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.001 |
| Science and technology studies | 0.001 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.001 |
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 teacher head, 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".