Yunsheng Zhang

Zhaoyang Zhang, Zhen Ren, Chao Tao et al.

Self-supervised contrastive learning (SSCL) has achieved significant milestones in remote sensing image (RSI) understanding. Its essence lies in designing an unsupervised instance discrimination pretext task to extract image features from a large number of unlabeled images that are beneficial for downstream tasks. However, existing instance discrimination based SSCL suffer from two limitations when applied to the RSI semantic segmentation task: 1) Positive sample confounding issue; 2) Feature adaptation bias. It introduces a feature adaptation bias when applied to semantic segmentation tasks that require pixel-level or object-level features. In this study, We observed that the discrimination information can be mapped to specific regions in RSI through the gradient of unsupervised contrastive loss, these specific regions tend to contain singular ground objects. Based on this, we propose contrastive learning with Gradient guided Sampling Strategy (GraSS) for RSI semantic segmentation. GraSS consists of two stages: Instance Discrimination warm-up (ID warm-up) and Gradient guided Sampling contrastive training (GS training). The ID warm-up aims to provide initial discrimination information to the contrastive loss gradients. The GS training stage aims to utilize the discrimination information contained in the contrastive loss gradients and adaptively select regions in RSI patches that contain more singular ground objects, in order to construct new positive and negative samples. Experimental results on three open datasets demonstrate that GraSS effectively enhances the performance of SSCL in high-resolution RSI semantic segmentation. Compared to seven baseline methods from five different types of SSCL, GraSS achieves an average improvement of 1.57\% and a maximum improvement of 3.58\% in terms of mean intersection over the union. The source code is available at https://github.com/GeoX-Lab/GraSS

Weirui Ye, Yunsheng Zhang, Haoyang Weng et al.

Reinforcement learning (RL) is a promising approach for solving robotic manipulation tasks. However, it is challenging to apply the RL algorithms directly in the real world. For one thing, RL is data-intensive and typically requires millions of interactions with environments, which are impractical in real scenarios. For another, it is necessary to make heavy engineering efforts to design reward functions manually. To address these issues, we leverage foundation models in this paper. We propose Reinforcement Learning with Foundation Priors (RLFP) to utilize guidance and feedback from policy, value, and success-reward foundation models. Within this framework, we introduce the Foundation-guided Actor-Critic (FAC) algorithm, which enables embodied agents to explore more efficiently with automatic reward functions. The benefits of our framework are threefold: (1) \textit{sample efficient}; (2) \textit{minimal and effective reward engineering}; (3) \textit{agnostic to foundation model forms and robust to noisy priors}. Our method achieves remarkable performances in various manipulation tasks on both real robots and in simulation. Across 5 dexterous tasks with real robots, FAC achieves an average success rate of 86\% after one hour of real-time learning. Across 8 tasks in the simulated Meta-world, FAC achieves 100\% success rates in 7/8 tasks under less than 100k frames (about 1-hour training), outperforming baseline methods with manual-designed rewards in 1M frames. We believe the RLFP framework can enable future robots to explore and learn autonomously in the physical world for more tasks. Visualizations and code are available at \url{https://yewr.github.io/rlfp}.

Run Shao, Zhaoyang Zhang, Chao Tao et al.

The tokenizer, as one of the fundamental components of large models, has long been overlooked or even misunderstood in visual tasks. One key factor of the great comprehension power of the large language model is that natural language tokenizers utilize meaningful words or subwords as the basic elements of language. In contrast, mainstream visual tokenizers, represented by patch-based methods such as Patch Embed, rely on meaningless rectangular patches as basic elements of vision, which cannot serve as effectively as words or subwords in language. Starting from the essence of the tokenizer, we defined semantically independent regions (SIRs) for vision. We designed a simple HOmogeneous visual tOKenizer: HOOK. HOOK mainly consists of two modules: the Object Perception Module (OPM) and the Object Vectorization Module (OVM). To achieve homogeneity, the OPM splits the image into 4*4 pixel seeds and then utilizes the attention mechanism to perceive SIRs. The OVM employs cross-attention to merge seeds within the same SIR. To achieve adaptability, the OVM defines a variable number of learnable vectors as cross-attention queries, allowing for the adjustment of token quantity. We conducted experiments on the NWPU-RESISC45, WHU-RS19 classification dataset, and GID5 segmentation dataset for sparse and dense tasks. The results demonstrate that the visual tokens obtained by HOOK correspond to individual objects, which demonstrates homogeneity. HOOK outperformed Patch Embed by 6\% and 10\% in the two tasks and achieved state-of-the-art performance compared to the baselines used for comparison. Compared to Patch Embed, which requires more than one hundred tokens for one image, HOOK requires only 6 and 8 tokens for sparse and dense tasks, respectively, resulting in efficiency improvements of 1.5 to 2.8 times. The code is available at https://github.com/GeoX-Lab/Hook.

Xiang Yang, Yongli Wang, HaiFeng Li et al.

Feed-forward 3D reconstruction has recently demonstrated strong generalization across diverse scenes, yet its performance in UAV imagery remains underexplored due to distinctive acquisition geometries, large viewpoint variations, and ambiguity between horizontal field of view and flight height. We present UAVFF3D, a geometry-aware benchmark for feed-forward UAV 3D reconstruction, comprising over 170K real UAV images and more than 370K high-quality synthetic images. The benchmark also includes a challenging diagnostic test subset designed to analyze model behavior under UAV-specific geometric ambiguities.Building on UAVFF3D, we propose an evaluation protocol that jointly assesses camera-geometry estimation and reconstruction accuracy, addressing limitations of existing evaluations that rely on separate alignments. Experiments on four representative feed-forward reconstruction models show that UAV-domain adaptation substantially improves performance, reducing Ray Error by up to 84.2%, Pose ATE by up to 76.0%, and Chamfer Distance by up to 41.1%. Further analysis reveals that domain adaptation mitigates rotation-estimation degradation in oblique-view scenes and improves robustness under horizontal-field-of-view/height ambiguity. Incorporating camera priors further enhances reconstruction performance under UAV-specific acquisition geometries.

Yunsheng Zhang, Jianguo Yao, Ruixiang Zhang et al.

Deep Neural Network (DNN) based point cloud semantic segmentation has presented significant achievements on large-scale labeled aerial laser point cloud datasets. However, annotating such large-scaled point clouds is time-consuming. Due to density variations and spatial heterogeneity of the Airborne Laser Scanning (ALS) point clouds, DNNs lack generalization capability and thus lead to unpromising semantic segmentation, as the DNN trained in one region underperform when directly utilized in other regions. However, Self-Supervised Learning (SSL) is a promising way to solve this problem by pre-training a DNN model utilizing unlabeled samples followed by a fine-tuned downstream task involving very limited labels. Hence, this work proposes a hard-negative sample aware self-supervised contrastive learning method to pre-train the model for semantic segmentation. The traditional contrastive learning for point clouds selects the hardest negative samples by solely relying on the distance between the embedded features derived from the learning process, potentially evolving some negative samples from the same classes to reduce the contrastive learning effectiveness. Therefore, we design an AbsPAN (Absolute Positive And Negative samples) strategy based on k-means clustering to filter the possible false-negative samples. Experiments on two typical ALS benchmark datasets demonstrate that the proposed method is more appealing than supervised training schemes without pre-training. Especially when the labels are severely inadequate (10% of the ISPRS training set), the results obtained by the proposed HAVANA method still exceed 94% of the supervised paradigm performance with full training set.

Yunsheng Zhang

We present the Incremental Generative Monte Carlo (IGMC) method, designed to measure uncertainty in deep neural networks using deep generative approaches. IGMC iteratively trains generative models, adding their output to the dataset, to compute the posterior distribution of the expectation of a random variable. We provide a theoretical guarantee of the convergence rate of IGMC relative to the sample size and sampling depth. Due to its compatibility with deep generative approaches, IGMC is adaptable to both neural network classification and regression tasks. We empirically study the behavior of IGMC on the MNIST digit classification task.