MS-ASL: A Large-Scale Data Set and Benchmark for Understanding American Sign Language

TL;DR

This paper introduces MS-ASL, a large-scale, publicly available dataset with over 25,000 videos for sign language recognition, and demonstrates the effectiveness of the I3D model in recognizing 1000 signs in real-world conditions.

Contribution

It provides the first large-scale, real-life sign language dataset with extensive annotations and evaluates state-of-the-art models, proposing I3D as a highly effective architecture.

Findings

I3D outperforms previous models significantly.

The dataset enables signer-independent recognition.

Recognition of 1000 signs in unconstrained conditions is feasible.

Abstract

Sign language recognition is a challenging and often underestimated problem comprising multi-modal articulators (handshape, orientation, movement, upper body and face) that integrate asynchronously on multiple streams. Learning powerful statistical models in such a scenario requires much data, particularly to apply recent advances of the field. However, labeled data is a scarce resource for sign language due to the enormous cost of transcribing these unwritten languages. We propose the first real-life large-scale sign language data set comprising over 25,000 annotated videos, which we thoroughly evaluate with state-of-the-art methods from sign and related action recognition. Unlike the current state-of-the-art, the data set allows to investigate the generalization to unseen individuals (signer-independent test) in a realistic setting with over 200 signers. Previous work mostly deals with limited vocabulary tasks, while here, we cover a large class count of 1000 signs in challenging and unconstrained real-life recording conditions. We further propose I3D, known from video classifications, as a powerful and suitable architecture for sign language recognition, outperforming the current state-of-the-art by a large margin. The data set is publicly available to the community.