TY - DATA T1 - Extension to IndustReal Dataset: Assembly State Recognition with Error States PY - 2024/10/28 AU - Tim J. Schoonbeek AU - Goutham Balachandran AU - Hans Onvlee AU - Tim Houben AU - Shao-Hsuan Hung AU - Jacek Kustra AU - Peter H.N. de With AU - Fons van der Sommen UR - DO - 10.4121/611adbc7-7935-43a6-8c3f-b2260a508e73.v1 KW - assembly state recognition KW - error recognition KW - assembly error KW - procedure error KW - computer vision KW - industrial AI KW - industry 4.0 KW - procedure understanding KW - egocentric N2 - <p>This dataset contains the assembly state images and annotations as an extension to the public <a href="https://data.4tu.nl/datasets/b008dd74-020d-4ea4-a8ba-7bb60769d224" target="_blank">IndustReal dataset</a>. Whereas the original dataset only classifies assembly errors as "error class", we define fine-grained annotations and error type classifications for every assembly error in the dataset. These labels enable researchers to continue studying procedural and execution errors in assembly and maintenance tasks.</p> ER -