Fei Gao

Chang Jiang, Fei Gao, Biao Ma et al.

We present a novel framework for exemplar based image translation. Recent advanced methods for this task mainly focus on establishing cross-domain semantic correspondence, which sequentially dominates image generation in the manner of local style control. Unfortunately, cross-domain semantic matching is challenging; and matching errors ultimately degrade the quality of generated images. To overcome this challenge, we improve the accuracy of matching on the one hand, and diminish the role of matching in image generation on the other hand. To achieve the former, we propose a masked and adaptive transformer (MAT) for learning accurate cross-domain correspondence, and executing context-aware feature augmentation. To achieve the latter, we use source features of the input and global style codes of the exemplar, as supplementary information, for decoding an image. Besides, we devise a novel contrastive style learning method, for acquire quality-discriminative style representations, which in turn benefit high-quality image generation. Experimental results show that our method, dubbed MATEBIT, performs considerably better than state-of-the-art methods, in diverse image translation tasks. The codes are available at \url{https://github.com/AiArt-HDU/MATEBIT}.

Ruibin Zhang, Donglai Xue, Yuhan Wang et al.

Millimeter wave (mmWave) radars have attracted significant attention from both academia and industry due to their capability to operate in extreme weather conditions. However, they face challenges in terms of sparsity and noise interference, which hinder their application in the field of micro aerial vehicle (MAV) autonomous navigation. To this end, this paper proposes a novel approach to dense and accurate mmWave radar point cloud construction via cross-modal learning. Specifically, we introduce diffusion models, which possess state-of-the-art performance in generative modeling, to predict LiDAR-like point clouds from paired raw radar data. We also incorporate the most recent diffusion model inference accelerating techniques to ensure that the proposed method can be implemented on MAVs with limited computing resources.We validate the proposed method through extensive benchmark comparisons and real-world experiments, demonstrating its superior performance and generalization ability. Code and pretrained models will be available at https://github.com/ZJU-FAST-Lab/Radar-Diffusion.

Fei Gao, Yifan Zhu, Chang Jiang et al.

Facial sketch synthesis (FSS) aims to generate a vivid sketch portrait from a given facial photo. Existing FSS methods merely rely on 2D representations of facial semantic or appearance. However, professional human artists usually use outlines or shadings to covey 3D geometry. Thus facial 3D geometry (e.g. depth map) is extremely important for FSS. Besides, different artists may use diverse drawing techniques and create multiple styles of sketches; but the style is globally consistent in a sketch. Inspired by such observations, in this paper, we propose a novel Human-Inspired Dynamic Adaptation (HIDA) method. Specially, we propose to dynamically modulate neuron activations based on a joint consideration of both facial 3D geometry and 2D appearance, as well as globally consistent style control. Besides, we use deformable convolutions at coarse-scales to align deep features, for generating abstract and distinct outlines. Experiments show that HIDA can generate high-quality sketches in multiple styles, and significantly outperforms previous methods, over a large range of challenging faces. Besides, HIDA allows precise style control of the synthesized sketch, and generalizes well to natural scenes and other artistic styles. Our code and results have been released online at: https://github.com/AiArt-HDU/HIDA.

Biao Ma, Fei Gao, Chang Jiang et al.

Neural radiance fields (NeRF) based methods have shown amazing performance in synthesizing 3D-consistent photographic images, but fail to generate multi-view portrait drawings. The key is that the basic assumption of these methods -- a surface point is consistent when rendered from different views -- doesn't hold for drawings. In a portrait drawing, the appearance of a facial point may changes when viewed from different angles. Besides, portrait drawings usually present little 3D information and suffer from insufficient training data. To combat this challenge, in this paper, we propose a Semantic-Aware GEnerator (SAGE) for synthesizing multi-view portrait drawings. Our motivation is that facial semantic labels are view-consistent and correlate with drawing techniques. We therefore propose to collaboratively synthesize multi-view semantic maps and the corresponding portrait drawings. To facilitate training, we design a semantic-aware domain translator, which generates portrait drawings based on features of photographic faces. In addition, use data augmentation via synthesis to mitigate collapsed results. We apply SAGE to synthesize multi-view portrait drawings in diverse artistic styles. Experimental results show that SAGE achieves significantly superior or highly competitive performance, compared to existing 3D-aware image synthesis methods. The codes are available at https://github.com/AiArt-HDU/SAGE.

Luxin Han, Fei Gao, Boyu Zhou et al.

Euclidean Signed Distance Field (ESDF) is useful for online motion planning of aerial robots since it can easily query the distance and gradient information against obstacles. Fast incrementally built ESDF map is the bottleneck for conducting real-time motion planning. In this paper, we investigate this problem and propose a mapping system called FIESTA to build global ESDF map incrementally. By introducing two independent updating queues for inserting and deleting obstacles separately, and using Indexing Data Structures and Doubly Linked Lists for map maintenance, our algorithm updates as few as possible nodes using a BFS framework. Our ESDF map has high computational performance and produces near-optimal results. We show our method outperforms other up-to-date methods in term of performance and accuracy by both theory and experiments. We integrate FIESTA into a completed quadrotor system and validate it by both simulation and onboard experiments. We release our method as open-source software for the community.

Guangyu Zhao, Tianyue Wu, Yeke Chen et al.

Animals learn to adapt speed of their movements to their capabilities and the environment they observe. Mobile robots should also demonstrate this ability to trade-off aggressiveness and safety for efficiently accomplishing tasks. The aim of this work is to endow flight vehicles with the ability of speed adaptation in prior unknown and partially observable cluttered environments. We propose a hierarchical learning and planning framework where we utilize both well-established methods of model-based trajectory generation and trial-and-error that comprehensively learns a policy to dynamically configure the speed constraint. Technically, we use online reinforcement learning to obtain the deployable policy. The statistical results in simulation demonstrate the advantages of our method over the constant speed constraint baselines and an alternative method in terms of flight efficiency and safety. In particular, the policy behaves perception awareness, which distinguish it from alternative approaches. By deploying the policy to hardware, we verify that these advantages can be brought to the real world.

Mingyu Yu, Wei Wang, Yanjie Wei et al.

Recent advancements in adversarial jailbreak attacks have exposed critical vulnerabilities in Large Language Models (LLMs), enabling the circumvention of alignment safeguards through increasingly sophisticated prompt manipulations. Based on our experiments, we found that the effectiveness of jailbreak strategies is influenced by the comprehension ability of the attacked LLM. Building on this insight, we propose a capability-aware Multi-Encryption Framework (MEF) for evaluating vulnerabilities in black-box LLMs. Specifically, MEF first categorizes the comprehension ability level of the LLM, then applies different strategies accordingly: For models with limited comprehension ability, MEF adopts the Fu+En1 strategy, which integrates layered semantic mutations with an encryption technique, more effectively contributing to evasion of the LLM's defenses at the input and inference stages. For models with strong comprehension ability, MEF uses a more complex Fu+En1+En2 strategy, in which additional dual-ended encryption techniques are applied to the LLM's responses, further contributing to evasion of the LLM's defenses at the output stage. Experimental results demonstrate the effectiveness of our approach, achieving attack success rates of 98.9% on GPT-4o (29 May 2025 release) and 99.8% on GPT-4.1 (8 July 2025 release). Our work contributes to a deeper understanding of the vulnerabilities in current LLM alignment mechanisms.

Fei Gao, Ming Hu, Zhiyu Xie et al.

With advancements in AI infrastructure and Trusted Execution Environment (TEE) technology, Federated Learning as a Service (FLaaS) through JointCloud Computing (JCC) is promising to break through the resource constraints caused by heterogeneous edge devices in the traditional Federated Learning (FL) paradigm. Specifically, with the protection from TEE, data owners can achieve efficient model training with high-performance AI services in the cloud. By providing additional FL services, cloud service providers can achieve collaborative learning among data owners. However, FLaaS still faces three challenges, i.e., i) low training performance caused by heterogeneous data among data owners, ii) high communication overhead among different clouds (i.e., data centers), and iii) lack of efficient resource scheduling strategies to balance training time and cost. To address these challenges, this paper presents a novel asynchronous FL approach named NebulaFL for collaborative model training among multiple clouds. To address data heterogeneity issues, NebulaFL adopts a version control-based asynchronous FL training scheme in each data center to balance training time among data owners. To reduce communication overhead, NebulaFL adopts a decentralized model rotation mechanism to achieve effective knowledge sharing among data centers. To balance training time and cost, NebulaFL integrates a reward-guided strategy for data owners selection and resource scheduling. The experimental results demonstrate that, compared to the state-of-the-art FL methods, NebulaFL can achieve up to 5.71\% accuracy improvement. In addition, NebulaFL can reduce up to 50% communication overhead and 61.94% costs under a target accuracy.

Fei Gao, Peng Geng, Jiaqi Guo et al.

We introduce ApolloRL, an open platform for research in reinforcement learning for autonomous driving. The platform provides a complete closed-loop pipeline with training, simulation, and evaluation components. It comes with 300 hours of real-world data in driving scenarios and popular baselines such as Proximal Policy Optimization (PPO) and Soft Actor-Critic (SAC) agents. We elaborate in this paper on the architecture and the environment defined in the platform. In addition, we discuss the performance of the baseline agents in the ApolloRL environment.

Ailing Cheng, Jiaojiao Ye, Fei Yang et al.

Aiming at the accurate and effective coaxiality measurement for twist drill with irregular surface, an optical measurement mechanism is proposed in this paper. First, A high-precision rotation instrument based on four core units is designed, which can obtain the 3-D point cloud data of full angle for the twist drill. Second, in the data processing stage, an improved robust Gaussian mixture model is established for accurate and rapid blade back segmentation. To improve measurement efficiency, a rapid reconstruction method of the twist drill axis based on orthogonal synthesis is provided to locate the axial position of the maximum deviation from the benchmark by utilizing the extracted blade back data. Finally, by calculating the maximum radial Euclidean distance from the benchmark, the coaxiality error of the twist drill is obtained. Comparing with other measurement methods, experimental results show that our proposed method is effective with high precision of 3 um and high efficiency of less than 3 s/pc. The result demonstrate that the proposed method is effective, robust and automatic, it can be applied in many actual industrial scene.

Lun Quan, Longji Yin, Chao Xu et al.

For aerial swarms, navigation in a prescribed formation is widely practiced in various scenarios. However, the associated planning strategies typically lack the capability of avoiding obstacles in cluttered environments. To address this deficiency, we present an optimization-based method that ensures collision-free trajectory generation for formation flight. In this paper, a novel differentiable metric is proposed to quantify the overall similarity distance between formations. We then formulate this metric into an optimization framework, which achieves spatial-temporal planning using polynomial trajectories. Minimization over collision penalty is also incorporated into the framework, so that formation preservation and obstacle avoidance can be handled simultaneously. To validate the efficiency of our method, we conduct benchmark comparisons with other cutting-edge works. Integrated with an autonomous distributed aerial swarm system, the proposed method demonstrates its efficiency and robustness in real-world experiments with obstacle-rich surroundings. We will release the source code for the reference of the community.

Jialin Ji, Neng Pan, Chao Xu et al.

This paper proposes Elastic Tracker, a flexible trajectory planning framework that can deal with challenging tracking tasks with guaranteed safety and visibility. Firstly, an object detection and intension-free motion prediction method is designed. Then an occlusion-aware path finding method is proposed to provide a proper topology. A smart safe flight corridor generation strategy is designed with the guiding path. An analytical occlusion cost is evaluated. Finally, an effective trajectory optimization approach enables to generate a spatio-temporal optimal trajectory within the resultant flight corridor. Particular formulations are designed to guarantee both safety and visibility, with all the above requirements optimized jointly. The experimental results show that our method works more robustly but with less computation than the existing methods, even in some challenging tracking tasks.

Qianhao Wang, Botao He, Zhiren Xun et al.

Gaze is an intuitive and direct way to represent the intentions of an individual. However, when it comes to assistive aerial teleoperation which aims to perform operators' intention, rare attention has been paid to gaze. Existing methods obtain intention directly from the remote controller (RC) input, which is inaccurate, unstable, and unfriendly to non-professional operators. Further, most teleoperation works do not consider environment perception which is vital to guarantee safety. In this paper, we present GPA-Teleoperation, a gaze enhanced perception-aware assistive teleoperation framework, which addresses the above issues systematically. We capture the intention utilizing gaze information, and generate a topological path matching it. Then we refine the path into a safe and feasible trajectory which simultaneously enhances the perception awareness to the environment operators are interested in. Additionally, the proposed method is integrated into a customized quadrotor system. Extensive challenging indoor and outdoor real-world experiments and benchmark comparisons verify that the proposed system is reliable, robust and applicable to even unskilled users. We will release the source code of our system to benefit related researches.

Ruibin Zhang, Yuze Wu, Lixian Zhang et al.

Terrestrial-aerial bimodal vehicles bloom in both academia and industry because they incorporate both the high mobility of aerial vehicles and the long endurance of ground vehicles. In this work, we present an autonomous and adaptive navigation framework to bring complete autonomy to this class of vehicles. The framework mainly includes 1) a hierarchical motion planner that generates safe and low-power terrestrial-aerial trajectories in unknown environments and 2) a unified motion controller which dynamically adjusts energy consumption in terrestrial locomotion. Extensive real-world experiments and benchmark comparisons are conducted on a customized robot platform to validate the proposed framework's robustness and performance. During the tests, the robot safely traverses complex environments with terrestrial-aerial integrated mobility, and achieves $7\times$ energy savings in terrestrial locomotion. Finally, we will release our code and hardware configuration for the reference of the community.

Qianhao Wang, Yuman Gao, Jialin Ji et al.

The visibility of targets determines performance and even success rate of various applications, such as active slam, exploration, and target tracking. Therefore, it is crucial to take the visibility of targets into explicit account in trajectory planning. In this paper, we propose a general metric for target visibility, considering observation distance and angle as well as occlusion effect. We formulate this metric into a differentiable visibility cost function, with which spatial trajectory and yaw can be jointly optimized. Furthermore, this visibility-aware trajectory optimization handles dynamic feasibility of position and yaw simultaneously. To validate that our method is practical and generic, we integrate it into a customized quadrotor tracking system. The experimental results show that our visibility-aware planner performs more robustly and observes targets better. In order to benefit related researches, we release our code to the public.

Yingjian Wang, Jialin Ji, Qianhao Wang et al.

In this paper, we introduce a complete system for autonomous flight of quadrotors in dynamic environments with onboard sensing. Extended from existing work, we develop an occlusion-aware dynamic perception method based on depth images, which classifies obstacles as dynamic and static. For representing generic dynamic environment, we model dynamic objects with moving ellipsoids and fuse static ones into an occupancy grid map. To achieve dynamic avoidance, we design a planning method composed of modified kinodynamic path searching and gradient-based optimization. The method leverages manually constructed gradients without maintaining a signed distance field (SDF), making the planning procedure finished in milliseconds. We integrate the above methods into a customized quadrotor system and thoroughly test it in realworld experiments, verifying its effective collision avoidance in dynamic environments.

Lizi Wang, Hongkai Ye, Qianhao Wang et al.

In autonomous navigation of mobile robots, sensors suffer from massive occlusion in cluttered environments, leaving significant amount of space unknown during planning. In practice, treating the unknown space in optimistic or pessimistic ways both set limitations on planning performance, thus aggressiveness and safety cannot be satisfied at the same time. However, humans can infer the exact shape of the obstacles from only partial observation and generate non-conservative trajectories that avoid possible collisions in occluded space. Mimicking human behavior, in this paper, we propose a method based on deep neural network to predict occupancy distribution of unknown space reliably. Specifically, the proposed method utilizes contextual information of environments and learns from prior knowledge to predict obstacle distributions in occluded space. We use unlabeled and no-ground-truth data to train our network and successfully apply it to real-time navigation in unseen environments without any refinement. Results show that our method leverages the performance of a kinodynamic planner by improving security with no reduction of speed in clustered environments.