3D instance segmentation is crucial for applications demanding comprehensive 3D scene understanding. Here, a novel method is introduced that simultaneously learns coefficients and prototypes. Employing an overcomplete sampling strategy, the method produces an overcomplete set of instance predictions, from which the optimal ones are selected through a Non-Maximum Suppression (NMS) algorithm during inference. The obtained prototypes are visualizable and interpretable. The method demonstrates superior performance on S3DIS-blocks, consistently outperforming existing methods in mRec and mPrec. Moreover, it operates 32.9\% faster than the state-of-the-art. Notably, with only 0.8\% of the total inference time, the method exhibits an over 20-fold reduction in the variance of inference time compared to existing methods. These attributes render the method well-suited for practical applications requiring both rapid inference and high reliability.
Royen, RD, Denis, L & Munteanu, A 2024, 'Joint prototype and coefficient prediction for 3D instance segmentation', Electronics Letters, vol. 60, no. 5, e13137. https://doi.org/10.1049/ell2.13137
Royen, R. D., Denis, L., & Munteanu, A. (2024). Joint prototype and coefficient prediction for 3D instance segmentation. Electronics Letters, 60(5), Article e13137. https://doi.org/10.1049/ell2.13137
@article{1523c6ae89a14d6cb952f85efa3548b7,
title = "Joint prototype and coefficient prediction for 3D instance segmentation",
abstract = "3D instance segmentation is crucial for applications demanding comprehensive 3D scene understanding. Here, a novel method is introduced that simultaneously learns coefficients and prototypes. Employing an overcomplete sampling strategy, the method produces an overcomplete set of instance predictions, from which the optimal ones are selected through a Non-Maximum Suppression (NMS) algorithm during inference. The obtained prototypes are visualizable and interpretable. The method demonstrates superior performance on S3DIS-blocks, consistently outperforming existing methods in mRec and mPrec. Moreover, it operates 32.9\% faster than the state-of-the-art. Notably, with only 0.8\% of the total inference time, the method exhibits an over 20-fold reduction in the variance of inference time compared to existing methods. These attributes render the method well-suited for practical applications requiring both rapid inference and high reliability.",
author = "Royen, \{Remco Donovan\} and Leon Denis and Adrian Munteanu",
note = "Publisher Copyright: {\textcopyright} 2024 The Authors. Electronics Letters published by John Wiley \& Sons Ltd on behalf of The Institution of Engineering and Technology.",
year = "2024",
month = mar,
doi = "10.1049/ell2.13137",
language = "English",
volume = "60",
journal = "Electronics Letters",
issn = "0013-5194",
publisher = "Institution of Engineering and Technology",
number = "5",
}