Zheng Liao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zheng Liao has a total of 8 co-authors. Here are the top co-authors by number of shared publications:
- Gang Li
School of Information Engineering
1 shared publication - Jin Li
School of Power and Energy, Northwestern Polytechnical University, Xi’an 710072, China
1 shared publication - Jingqi Qiao1 shared publication
- Shuai Ma
Shenyang Engine Research Institute, Shenyang 110066, China
1 shared publication - Sijie Liu
School of Power and Energy, Northwestern Polytechnical University, Xi’an 710072, China
1 shared publication Tianwei Wang
1 shared publication
Latest Publications
- Open Access
- 4 October 2026
- Published by Elsevier in SSRN Electronic Journal
Sparse-Aware Geometric Tokenization with Modality-Specific Reconstruction for Front-View RGB–LiDAR Semantic Segmentation
Front-view RGB--LiDAR semantic segmentation requires the joint exploitation of dense visual appearance and sparse geometric observations within a unified image plane. However, projected LiDAR is characterized by large empty regions, discontinuous local responses, and heterogeneous geometric-variable distributions, making standard patch embedding designed for dense image signals suboptimal for such non-uniform sparse inputs. To address this issue, we propose a sparse-aware geometric tokenization framework for front-view RGB--LiDAR semantic segmentation. Specifically, a Sparse-Aware Geometric Embedding (SAGE) module organizes three-dimensional Cartesian coordinates together with deterministically derived range and angular variables into a six-channel geometric representation, and integrates local response aggregation, token normalization, and empty-patch initial-response suppression into the LiDAR token construction process. Within a dual-stream Transformer architecture, modality-specific hierarchical reconstruction is further employed to maintain independent token reassembly and spatial recovery paths for RGB and LiDAR, thereby preserving modality-specific token-to-pixel propagation before cross-modal fusion. A lightweight Cross-modal Locality-Preserving Refinement (CLPR) module is additionally introduced before scale-wise fusion to refine local LiDAR representations using aligned RGB context. Experimental results show that introducing the hierarchical reconstruction framework improves Vehicle IoU and Human IoU from 91.33\% and 66.91\% to 94.03\% and 76.38\%, respectively. Incorporating SAGE further increases the two metrics to 95.07\% and 79.10\%, while the complete model achieves 95.09\% Vehicle IoU and 79.61\% Human IoU. Additional analyses demonstrate that SAGE consistently yields positive gains across different LiDAR geometric-support levels and road-recording conditions, whereas the contribution of CLPR is mainly concentrated on a subset of challenging Human boundary regions. These results indicate that sparse-aware geometric tokenization can effectively complement hierarchical dual-stream spatial reconstruction and improve pixel-level RGB--LiDAR semantic segmentation in front-view driving scenes. The code will be released upon acceptance.
- 4 October 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessSparse-Aware Geometric Tokenization with Modality-Specific Reconstruction for Front-View RGB–LiDAR Semantic Segmentation
Front-view RGB--LiDAR semantic segmentation requires the joint exploitation of dense visual appearance and sparse geometric observations within a unified image plane. However, projected LiDAR is characterized by large empty regions, discontinuous local responses, and heterogeneous geometric-variable distributions, making standard patch embedding designed for dense image signals suboptimal for such non-uniform sparse inputs. To address this issue, we propose a sparse-aware geometric tokenization framework for front-view RGB--LiDAR semantic segmentation. Specifically, a Sparse-Aware Geometric Embedding (SAGE) module organizes three-dimensional Cartesian coordinates together with deterministically derived range and angular variables into a six-channel geometric representation, and integrates local response aggregation, token normalization, and empty-patch initial-response suppression into the LiDAR token construction process. Within a dual-stream Transformer architecture, modality-specific hierarchical reconstruction is further employed to maintain independent token reassembly and spatial recovery paths for RGB and LiDAR, thereby preserving modality-specific token-to-pixel propagation before cross-modal fusion. A lightweight Cross-modal Locality-Preserving Refinement (CLPR) module is additionally introduced before scale-wise fusion to refine local LiDAR representations using aligned RGB context. Experimental results show that introducing the hierarchical reconstruction framework improves Vehicle IoU and Human IoU from 91.33\% and 66.91\% to 94.03\% and 76.38\%, respectively. Incorporating SAGE further increases the two metrics to 95.07\% and 79.10\%, while the complete model achieves 95.09\% Vehicle IoU and 79.61\% Human IoU. Additional analyses demonstrate that SAGE consistently yields positive gains across different LiDAR geometric-support levels and road-recording conditions, whereas the contribution of CLPR is mainly concentrated on a subset of challenging Human boundary regions. These results indicate that sparse-aware geometric tokenization can effectively complement hierarchical dual-stream spatial reconstruction and improve pixel-level RGB--LiDAR semantic segmentation in front-view driving scenes. The code will be released upon acceptance.
- Open Access
Aero-Engine Remain Useful Life Estimation via Scope-Coordinated Attention-Based Network
Research on the assessment of the remaining useful life (RUL) has garnered significant attention because of its critical relevance in prognostics and health management (PHM) across various sectors. Recently, data-driven methodologies have become increasingly important for RUL prediction. However, these methods often struggle to capture long-term dependencies and possess a limited receptive field, restricting their effectiveness in various RUL prediction scenarios. To address these limitations, this study proposes a novel approach called the scope-coordinated attention-based (SCAB) network for RUL prediction. The initial design features a novel multichannel feature integration block, which aims to effectively capture and integrate essential information from raw sensor data. Additionally, it is designed to expand the receptive field by capturing rich and diverse features for improved representation. Subsequently, a dual-attention block refines information and further expands the receptive field in both the channel and spatial domain. Moreover, a feature pyramid block with adaptive self-attention is developed to effectively capture long-term dependencies, further enhancing the information’s detail and features by the multiscale feature fusion mechanism. The efficacy of the proposed SCAB model for RUL estimation was validated using the C-MAPSS public dataset. In comparison experiments, the SCAB model outperformed other methods in the FD002 subset while demonstrating excellent performances in FD001, FD003, and FD004. These results confirm that the SCAB model exhibits robust and superior performance in RUL prediction across various aeroengine scenarios.
Open AccessAero-Engine Remain Useful Life Estimation via Scope-Coordinated Attention-Based Network
Research on the assessment of the remaining useful life (RUL) has garnered significant attention because of its critical relevance in prognostics and health management (PHM) across various sectors. Recently, data-driven methodologies have become increasingly important for RUL prediction. However, these methods often struggle to capture long-term dependencies and possess a limited receptive field, restricting their effectiveness in various RUL prediction scenarios. To address these limitations, this study proposes a novel approach called the scope-coordinated attention-based (SCAB) network for RUL prediction. The initial design features a novel multichannel feature integration block, which aims to effectively capture and integrate essential information from raw sensor data. Additionally, it is designed to expand the receptive field by capturing rich and diverse features for improved representation. Subsequently, a dual-attention block refines information and further expands the receptive field in both the channel and spatial domain. Moreover, a feature pyramid block with adaptive self-attention is developed to effectively capture long-term dependencies, further enhancing the information’s detail and features by the multiscale feature fusion mechanism. The efficacy of the proposed SCAB model for RUL estimation was validated using the C-MAPSS public dataset. In comparison experiments, the SCAB model outperformed other methods in the FD002 subset while demonstrating excellent performances in FD001, FD003, and FD004. These results confirm that the SCAB model exhibits robust and superior performance in RUL prediction across various aeroengine scenarios.