AUTSL stands for Ankara University Turkish Sign Language Dataset, which is a large-scale multimodal dataset containing independent Turkish sign language videos. The dataset contains 226 gestures performed by 43 different signers, with a total of 38,336 video samples.

The samples were recorded with a Microsoft Kinect v2 in RGB, depth, and skeleton formats. The resolution of the videos is 512×512 pixels. The skeleton data contains the spatial coordinates of 25 nodes on the signer's body, i.e. (x, y), and the nodes are aligned with the 512×512 data.