arXiv subjects

Xingzhao Liu

Publications and source records attributed to Xingzhao Liu.

3 recordsLinked to original sources

arXiv2026

AirSpatialBot: A Spatially-Aware Aerial Agent for Fine-Grained Vehicle Attribute Recognization and Retrieval

Despite notable advancements in remote sensing vision-language models (VLMs), existing models often struggle with spatial understanding, limiting their effectiveness in real-world applications. To push the boundaries of VLMs in remote sensing, we specifically address vehicle imagery captured by drones and introduce a spatially-aware dataset AirSpatial, which comprises over 206K instructions and introduces two novel tasks: Spatial Grounding and Spatial Question Answering. It is also the first remote sensing grounding dataset to provide 3DBB. To effectively leverage existing image understanding of VLMs to spatial domains, we adopt a two-stage training strategy comprising Image Understanding Pre-training and Spatial Understanding Fine-tuning. Utilizing this trained spatially-aware VLM, we develop an aerial agent, AirSpatialBot, which is capable of fine-grained vehicle attribute recognition and retrieval. By dynamically integrating task planning, image understanding, spatial understanding, and task execution capabilities, AirSpatialBot adapts to diverse query requirements. Experimental results validate the effectiveness of our approach, revealing the spatial limitations of existing VLMs while providing valuable insights. The model, code, and datasets will be released at https://github.com/VisionXLab/AirSpatialBot

cs.CV↗

arXiv2024

Robust Wideband Channel Estimation for mmWave Massive MIMO Systems With Beam Squint

This paper investigates the robust wideband channel estimation problem in the millimeter-wave (mmWave) massive multiple-input multiple-output (MIMO) systems. In such a scenario, the beam squint effect that the array response vectors vary with different frequencies and the impulsive noise are in existence, which pose great challenges for accurate channel estimation. Directly applying the existing channel estimation methods usually suffers significant performance degradation, since they are proposed based on the assumptions of frequency-invariant array response vectors and Gaussian distributed noise. To address these issues, this paper proposes a novel wideband channel estimation method with robustness to impulsive noise. Specifically, the proposed method incorporates a cyclic refinement step to overcome the estimation inaccuracy caused by the greedy nature of matching-pursuit-esque algorithms. In particular, the generalized $\ell_p$-norm minimization criterion is adopted in Newton's method to improve the performance robustness against the non-Gaussian impulsive noise. Numerical results are provided to verify the superior performance of the proposed method over the existing representative benchmarks.

eess.SP↗

arXiv2022

MMRotate: A Rotated Object Detection Benchmark using PyTorch

We present an open-source toolbox, named MMRotate, which provides a coherent algorithm framework of training, inferring, and evaluation for the popular rotated object detection algorithm based on deep learning. MMRotate implements 18 state-of-the-art algorithms and supports the three most frequently used angle definition methods. To facilitate future research and industrial applications of rotated object detection-related problems, we also provide a large number of trained models and detailed benchmarks to give insights into the performance of rotated object detection. MMRotate is publicly released at https://github.com/open-mmlab/mmrotate.

cs.CV↗