Summary of Skelcap: Automated Generation Of Descriptive Text From Skeleton Keypoint Sequences, by Ali Emre Keskin and Hacer Yalim Keles
SkelCap: Automated Generation of Descriptive Text from Skeleton Keypoint Sequences
by Ali Emre Keskin, Hacer Yalim Keles
First submitted to arxiv on: 5 May 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary A new approach to addressing limitations in sign language datasets is proposed, focusing on textually describing body movements from skeleton keypoint sequences. A comprehensive isolated Turkish sign language dataset (AUTSL) serves as a foundation for this effort. A baseline model, SkelCap, is developed for generating textual descriptions of body movements, processing skeleton keypoints data as a vector and utilizing a transformer neural network for sequence-to-sequence modeling. The model achieves promising results in extensive evaluations, including signer-agnostic and sign-agnostic assessments, with ROUGE-L and BLEU-4 scores of 0.98 and 0.94 respectively. This work aims to facilitate the creation of diverse sign language datasets, addressing challenges associated with gathering a varied group of signers. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Sign language has thousands of signs used globally, but most datasets cover only a limited selection. Creating diverse datasets is expensive and challenging because it requires gathering many different signers. Researchers developed a new way to create sign language datasets by describing body movements using skeleton keypoint sequences. They made a comprehensive dataset for isolated Turkish signs (AUTSL) and a model that generates descriptions of body movements (SkelCap). The model works well, scoring 0.98 and 0.94 in tests. |
Keywords
» Artificial intelligence » Bleu » Neural network » Rouge » Transformer
