# Let's Dance: Learning From Online Dance Videos

arXiv: https://arxiv.org/abs/1801.07388 · HTML: https://mlsanity.com/paper/1801.07388/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 23 Jan 2018.

## Built on by (most cited first, 7 of 7)

- [Unsupervised 3D Pose Estimation for Hierarchical Dance Video Recognition](https://mlsanity.com/paper/2109.09166/) (2021)
- [Fine-Grained Sports, Yoga, and Dance Postures Recognition: A Benchmark Analysis](https://mlsanity.com/paper/2308.00323/) (2023)
- [MoSA: Motion-Coherent Human Video Generation via Structure-Appearance Decoupling](https://mlsanity.com/paper/2508.17404/) (2025)
- [MOSA: Music Motion with Semantic Annotation Dataset for Cross-Modal Music Processing](https://mlsanity.com/paper/2406.06375/) (2024)
- [AnthroTAP: Learning Point Tracking with Real-World Motion](https://mlsanity.com/paper/2507.06233/) (2025)
- [Learning Music-Dance Representations through Explicit-Implicit Rhythm Synchronization](https://mlsanity.com/paper/2207.03190/) (2022)
- [Learning Quantised Structure-Preserving Motion Representations for Dance Fingerprinting](https://mlsanity.com/paper/2604.00927/) (2026)

## Abstract (arXiv)

In recent years, deep neural network approaches have naturally extended to the video domain, in their simplest case by aggregating per-frame classifications as a baseline for action recognition. A majority of the work in this area extends from the imaging domain, leading to visual-feature heavy approaches on temporal data. To address this issue we introduce "Let's Dance", a 1000 video dataset (and growing) comprised of 10 visually overlapping dance categories that require motion for their classification. We stress the important of human motion as a key distinguisher in our work given that, as we show in this work, visual information is not sufficient to classify motion-heavy categories. We compare our datasets' performance using imaging techniques with UCF-101 and demonstrate this inherent difficulty. We present a comparison of numerous state-of-the-art techniques on our dataset using three different representations (video, optical flow and multi-person pose data) in order to analyze these approaches. We discuss the motion parameterization of each of them and their value in learning to categorize online dance videos. Lastly, we release this dataset (and its three representations) for the research community to use.