Efficient LiDAR Reflectance Compression via Scanning Serialization

Jiahao Zhu, Kang You, Dandan Ding, Zhan Ma
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:80330-80346, 2025.

Abstract

Reflectance attributes in LiDAR point clouds provide essential information for downstream tasks but remain underexplored in neural compression methods. To address this, we introduce SerLiC, a serialization-based neural compression framework to fully exploit the intrinsic characteristics of LiDAR reflectance. SerLiC first transforms 3D LiDAR point clouds into 1D sequences via scan-order serialization, offering a device-centric perspective for reflectance analysis. Each point is then tokenized into a contextual representation comprising its sensor scanning index, radial distance, and prior reflectance, for effective dependencies exploration. For efficient sequential modeling, Mamba is incorporated with a dual parallelization scheme, enabling simultaneous autoregressive dependency capture and fast processing. Extensive experiments demonstrate that SerLiC attains over 2$\times$ volume reduction against the original reflectance data, outperforming the state-of-the-art method by up to 22% reduction of compressed bits while using only 2% of its parameters. Moreover, a lightweight version of SerLiC achieves $\geq 10$ fps (frames per second) with just 111K parameters, which is attractive for real applications.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-zhu25aa, title = {Efficient {L}i{DAR} Reflectance Compression via Scanning Serialization}, author = {Zhu, Jiahao and You, Kang and Ding, Dandan and Ma, Zhan}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {80330--80346}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/zhu25aa/zhu25aa.pdf}, url = {https://proceedings.mlr.press/v267/zhu25aa.html}, abstract = {Reflectance attributes in LiDAR point clouds provide essential information for downstream tasks but remain underexplored in neural compression methods. To address this, we introduce SerLiC, a serialization-based neural compression framework to fully exploit the intrinsic characteristics of LiDAR reflectance. SerLiC first transforms 3D LiDAR point clouds into 1D sequences via scan-order serialization, offering a device-centric perspective for reflectance analysis. Each point is then tokenized into a contextual representation comprising its sensor scanning index, radial distance, and prior reflectance, for effective dependencies exploration. For efficient sequential modeling, Mamba is incorporated with a dual parallelization scheme, enabling simultaneous autoregressive dependency capture and fast processing. Extensive experiments demonstrate that SerLiC attains over 2$\times$ volume reduction against the original reflectance data, outperforming the state-of-the-art method by up to 22% reduction of compressed bits while using only 2% of its parameters. Moreover, a lightweight version of SerLiC achieves $\geq 10$ fps (frames per second) with just 111K parameters, which is attractive for real applications.} }
Endnote
%0 Conference Paper %T Efficient LiDAR Reflectance Compression via Scanning Serialization %A Jiahao Zhu %A Kang You %A Dandan Ding %A Zhan Ma %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-zhu25aa %I PMLR %P 80330--80346 %U https://proceedings.mlr.press/v267/zhu25aa.html %V 267 %X Reflectance attributes in LiDAR point clouds provide essential information for downstream tasks but remain underexplored in neural compression methods. To address this, we introduce SerLiC, a serialization-based neural compression framework to fully exploit the intrinsic characteristics of LiDAR reflectance. SerLiC first transforms 3D LiDAR point clouds into 1D sequences via scan-order serialization, offering a device-centric perspective for reflectance analysis. Each point is then tokenized into a contextual representation comprising its sensor scanning index, radial distance, and prior reflectance, for effective dependencies exploration. For efficient sequential modeling, Mamba is incorporated with a dual parallelization scheme, enabling simultaneous autoregressive dependency capture and fast processing. Extensive experiments demonstrate that SerLiC attains over 2$\times$ volume reduction against the original reflectance data, outperforming the state-of-the-art method by up to 22% reduction of compressed bits while using only 2% of its parameters. Moreover, a lightweight version of SerLiC achieves $\geq 10$ fps (frames per second) with just 111K parameters, which is attractive for real applications.
APA
Zhu, J., You, K., Ding, D. & Ma, Z.. (2025). Efficient LiDAR Reflectance Compression via Scanning Serialization. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:80330-80346 Available from https://proceedings.mlr.press/v267/zhu25aa.html.

Related Material