Dong Han

Dong Han, Jihye Moon, Luís Roberto Mercado Díaz et al.

Most deep learning models of multiclass arrhythmia classification are tested on fingertip photoplethysmographic (PPG) data, which has higher signal-to-noise ratios compared to smartwatch-derived PPG, and the best reported sensitivity value for premature atrial/ventricular contraction (PAC/PVC) detection is only 75%. To improve upon PAC/PVC detection sensitivity while maintaining high AF detection, we use multi-modal data which incorporates 1D PPG, accelerometers, and heart rate data as the inputs to a computationally efficient 1D bi-directional Gated Recurrent Unit (1D-Bi-GRU) model to detect three arrhythmia classes. We used motion-artifact prone smartwatch PPG data from the NIH-funded Pulsewatch clinical trial. Our multimodal model tested on 72 subjects achieved an unprecedented 83% sensitivity for PAC/PVC detection while maintaining a high accuracy of 97.31% for AF detection. These results outperformed the best state-of-the-art model by 20.81% for PAC/PVC and 2.55% for AF detection even while our model was computationally more efficient (14 times lighter and 2.7 faster).

Dong Han, Yufan Jiang, Yong Li et al.

In this work, we leverage the pure skin color patch from the face image as the additional information to train an auxiliary skin color feature extractor and face recognition model in parallel to improve performance of state-of-the-art (SOTA) privacy-preserving face recognition (PPFR) systems. Our solution is robust against black-box attacking and well-established generative adversarial network (GAN) based image restoration. We analyze the potential risk in previous work, where the proposed cosine similarity computation might directly leak the protected precomputed embedding stored on the server side. We propose a Function Secret Sharing (FSS) based face embedding comparison protocol without any intermediate result leakage. In addition, we show in experiments that the proposed protocol is more efficient compared to the Secret Sharing (SS) based protocol.

Weida Wang, Dongchen Huang, Jiatong Li et al.

We introduce CMPhysBench, designed to assess the proficiency of Large Language Models (LLMs) in Condensed Matter Physics, as a novel Benchmark. CMPhysBench is composed of more than 520 graduate-level meticulously curated questions covering both representative subfields and foundational theoretical frameworks of condensed matter physics, such as magnetism, superconductivity, strongly correlated systems, etc. To ensure a deep understanding of the problem-solving process,we focus exclusively on calculation problems, requiring LLMs to independently generate comprehensive solutions. Meanwhile, leveraging tree-based representations of expressions, we introduce the Scalable Expression Edit Distance (SEED) score, which provides fine-grained (non-binary) partial credit and yields a more accurate assessment of similarity between prediction and ground-truth. Our results show that even the best models, Grok-4, reach only 36 average SEED score and 28% accuracy on CMPhysBench, underscoring a significant capability gap, especially for this practical and frontier domain relative to traditional physics. The code anddataset are publicly available at https://github.com/CMPhysBench/CMPhysBench.

Dong Han, Yong Li, Joachim Denzler

With the advancement of face recognition (FR) systems, privacy-preserving face recognition (PPFR) systems have gained popularity for their accurate recognition, enhanced facial privacy protection, and robustness to various attacks. However, there are limited studies to further verify privacy risks by reconstructing realistic high-resolution face images from embeddings of these systems, especially for PPFR. In this work, we propose the face embedding mapping (FEM), a general framework that explores Kolmogorov-Arnold Network (KAN) for conducting the embedding-to-face attack by leveraging pre-trained Identity-Preserving diffusion model against state-of-the-art (SOTA) FR and PPFR systems. Based on extensive experiments, we verify that reconstructed faces can be used for accessing other real-word FR systems. Besides, the proposed method shows the robustness in reconstructing faces from the partial and protected face embeddings. Moreover, FEM can be utilized as a tool for evaluating safety of FR and PPFR systems in terms of privacy leakage. All images used in this work are from public datasets.

Bing Han, Chen Zhu, Dong Han et al.

Vision-driven field monitoring is central to digital agriculture, yet models built on general-domain pretrained backbones often fail to generalize across tasks, owing to the interaction of fine, variable canopy structures with fluctuating field conditions. We present FoMo4Wheat, one of the first crop-domain vision foundation model pretrained with self-supervision on ImAg4Wheat, the largest and most diverse wheat image dataset to date (2.5 million high-resolution images collected over a decade at 30 global sites, spanning >2,000 genotypes and >500 environmental conditions). This wheat-specific pretraining yields representations that are robust for wheat and transferable to other crops and weeds. Across ten in-field vision tasks at canopy and organ levels, FoMo4Wheat models consistently outperform state-of-the-art models pretrained on general-domain dataset. These results demonstrate the value of crop-specific foundation models for reliable in-field perception and chart a path toward a universal crop foundation model with cross-species and cross-task capabilities. FoMo4Wheat models and the ImAg4Wheat dataset are publicly available online: https://github.com/PheniX-Lab/FoMo4Wheat and https://huggingface.co/PheniX-Lab/FoMo4Wheat. The demonstration website is: https://fomo4wheat.phenix-lab.com/.

Wei Chu, Yuanzhe Dong, Ke Tan et al.

OleSpeech-IV dataset is a large-scale multispeaker and multilingual conversational speech dataset with diverse topics. The audio content comes from publicly-available English podcasts, talk shows, teleconferences, and other conversations. Speaker names, turns, and transcripts are human-sourced and refined by a proprietary pipeline, while additional information such as timestamps and confidence scores is derived from the pipeline. The IV denotes its position as Tier IV in the Olewave dataset series. In addition, we have open-sourced a subset, OleSpeech-IV-2025-EN-AR-100, for non-commercial research use.

Salaheldin Mohamed, Dong Han, Yong Li

Text-to-image (T2I) models have significantly advanced the development of artificial intelligence, enabling the generation of high-quality images in diverse contexts based on specific text prompts. However, existing T2I-based methods often struggle to accurately reproduce the appearance of individuals from a reference image and to create novel representations of those individuals in various settings. To address this, we leverage the pre-trained UNet from Stable Diffusion to incorporate the target face image directly into the generation process. Our approach diverges from prior methods that depend on fixed encoders or static face embeddings, which often fail to bridge encoding gaps. Instead, we capitalize on UNet's sophisticated encoding capabilities to process reference images across multiple scales. By innovatively altering the cross-attention layers of the UNet, we effectively fuse individual identities into the generative process. This strategic integration of facial features across various scales not only enhances the robustness and consistency of the generated images but also facilitates efficient multi-reference and multi-identity generation. Our method sets a new benchmark in identity-preserving image generation, delivering state-of-the-art results in similarity metrics while maintaining prompt alignment.

Dong Han, Yong Li, Joachim Denzler

With the advancement of face reconstruction (FR) systems, privacy-preserving face recognition (PPFR) has gained popularity for its secure face recognition, enhanced facial privacy protection, and robustness to various attacks. Besides, specific models and algorithms are proposed for face embedding protection by mapping embeddings to a secure space. However, there is a lack of studies on investigating and evaluating the possibility of extracting face images from embeddings of those systems, especially for PPFR. In this work, we introduce the first approach to exploit Kolmogorov-Arnold Network (KAN) for conducting embedding-to-face attacks against state-of-the-art (SOTA) FR and PPFR systems. Face embedding mapping (FEM) models are proposed to learn the distribution mapping relation between the embeddings from the initial domain and target domain. In comparison with Multi-Layer Perceptrons (MLP), we provide two variants, FEM-KAN and FEM-MLP, for efficient non-linear embedding-to-embedding mapping in order to reconstruct realistic face images from the corresponding face embedding. To verify our methods, we conduct extensive experiments with various PPFR and FR models. We also measure reconstructed face images with different metrics to evaluate the image quality. Through comprehensive experiments, we demonstrate the effectiveness of FEMs in accurate embedding mapping and face reconstruction.

Enhui Ma, Lijun Zhou, Tao Tang et al.

End-to-end planning methods are the de facto standard of the current autonomous driving system, while the robustness of the data-driven approaches suffers due to the notorious long-tail problem (i.e., rare but safety-critical failure cases). In this work, we explore whether recent diffusion-based video generation methods (a.k.a. world models), paired with structured 3D layouts, can enable a fully automated pipeline to self-correct such failure cases. We first introduce an agent to simulate the role of product manager, dubbed PM-Agent, which formulates data requirements to collect data similar to the failure cases. Then, we use a generative model that can simulate both data collection and annotation. However, existing generative models struggle to generate high-fidelity data conditioned on 3D layouts. To address this, we propose DriveSora, which can generate spatiotemporally consistent videos aligned with the 3D annotations requested by PM-Agent. We integrate these components into our self-correcting agentic system, CorrectAD. Importantly, our pipeline is an end-to-end model-agnostic and can be applied to improve any end-to-end planner. Evaluated on both nuScenes and a more challenging in-house dataset across multiple end-to-end planners, CorrectAD corrects 62.5% and 49.8% of failure cases, reducing collision rates by 39% and 27%, respectively.

Dong Han, Zhehong Ai, Pengxiang Cai et al.

Bayesian optimization (BO) is a powerful tool for scientific discovery in chemistry, yet its efficiency is often hampered by the sparse experimental data and vast search space. Here, we introduce ChemBOMAS: a large language model (LLM)-enhanced multi-agent system that accelerates BO through synergistic data- and knowledge-driven strategies. Firstly, the data-driven strategy involves an 8B-scale LLM regressor fine-tuned on a mere 1% labeled samples for pseudo-data generation, robustly initializing the optimization process. Secondly, the knowledge-driven strategy employs a hybrid Retrieval-Augmented Generation approach to guide LLM in dividing the search space while mitigating LLM hallucinations. An Upper Confidence Bound algorithm then identifies high-potential subspaces within this established partition. Across the LLM-refined subspaces and supported by LLM-generated data, BO achieves the improvement of effectiveness and efficiency. Comprehensive evaluations across multiple scientific benchmarks demonstrate that ChemBOMAS set a new state-of-the-art, accelerating optimization efficiency by up to 5-fold compared to baseline methods.

Enhui Ma, Lijun Zhou, Tao Tang et al.

Using generative models to synthesize new data has become a de-facto standard in autonomous driving to address the data scarcity issue. Though existing approaches are able to boost perception models, we discover that these approaches fail to improve the performance of planning of end-to-end autonomous driving models as the generated videos are usually less than 8 frames and the spatial and temporal inconsistencies are not negligible. To this end, we propose Delphi, a novel diffusion-based long video generation method with a shared noise modeling mechanism across the multi-views to increase spatial consistency, and a feature-aligned module to achieves both precise controllability and temporal consistency. Our method can generate up to 40 frames of video without loss of consistency which is about 5 times longer compared with state-of-the-art methods. Instead of randomly generating new data, we further design a sampling policy to let Delphi generate new data that are similar to those failure cases to improve the sample efficiency. This is achieved by building a failure-case driven framework with the help of pre-trained visual language models. Our extensive experiment demonstrates that our Delphi generates a higher quality of long videos surpassing previous state-of-the-art methods. Consequentially, with only generating 4% of the training dataset size, our framework is able to go beyond perception and prediction tasks, for the first time to the best of our knowledge, boost the planning performance of the end-to-end autonomous driving model by a margin of 25%.

Dong Han, Yong Li, Joachim Denzler

Face recognition technology has been deployed in various real-life applications. The most sophisticated deep learning-based face recognition systems rely on training millions of face images through complex deep neural networks to achieve high accuracy. It is quite common for clients to upload face images to the service provider in order to access the model inference. However, the face image is a type of sensitive biometric attribute tied to the identity information of each user. Directly exposing the raw face image to the service provider poses a threat to the user's privacy. Current privacy-preserving approaches to face recognition focus on either concealing visual information on model input or protecting model output face embedding. The noticeable drop in recognition accuracy is a pitfall for most methods. This paper proposes a hybrid frequency-color fusion approach to reduce the input dimensionality of face recognition in the frequency domain. Moreover, sparse color information is also introduced to alleviate significant accuracy degradation after adding differential privacy noise. Besides, an identity-specific embedding mapping scheme is applied to protect original face embedding by enlarging the distance among identities. Lastly, secure multiparty computation is implemented for safely computing the embedding distance during model inference. The proposed method performs well on multiple widely used verification datasets. Moreover, it has around 2.6% to 4.2% higher accuracy than the state-of-the-art in the 1:N verification scenario.

Tianyu Liu, Qinghai Liao, Lu Gan et al.

Since early 2020, the coronavirus disease 2019 (COVID-19) has spread rapidly across the world. As at the date of writing this article, the disease has been globally reported in 223 countries and regions, infected over 108 million people and caused over 2.4 million deaths (https://covid19.who.int/, accessed on Feb. 17, 2021). Avoiding person-to-person transmission is an effective approach to control and prevent the pandemic. However, many daily activities, such as transporting goods in our daily life, inevitably involve person-to-person contact. Using an autonomous logistic vehicle to achieve contact-less goods transportation could alleviate this issue. For example, it can reduce the risk of virus transmission between the driver and customers. Moreover, many countries have imposed tough lockdown measures to reduce the virus transmission (e.g., retail, catering) during the pandemic, which causes inconveniences for human daily life. Autonomous vehicle can deliver the goods bought by humans, so that humans can get the goods without going out. These demands motivate us to develop an autonomous vehicle, named as Hercules, for contact-less goods transportation during the COVID-19 pandemic. The vehicle is evaluated through real-world delivering tasks under various traffic conditions.

Dong Han, Zuhao Zou, Lujia Wang et al.

In complex environments, low-cost and robust localization is a challenging problem. For example, in a GPSdenied environment, LiDAR can provide accurate position information, but the cost is high. In general, visual SLAM based localization methods become unreliable when the sunlight changes greatly. Therefore, inexpensive and reliable methods are required. In this paper, we propose a stereo visual localization method based on the prior LiDAR map. Different from the conventional visual localization system, we design a novel visual optimization model by matching planar information between the LiDAR map and visual image. Bundle adjustment is built by using coplanarity constraints. To solve the optimization problem, we use a graph-based optimization algorithm and a local window optimization method. Finally, we estimate a full six degrees of freedom (DOF) pose without scale drift. To validate the efficiency, the proposed method has been tested on the KITTI dataset. The results show that our method is more robust and accurate than the state-of-art ORB-SLAM2.

Yilong Zhu, Dong Han, Bohuan Xue et al.

Road curb detection is very important and necessary for autonomous driving because it can improve the safety and robustness of robot navigation in the outdoor environment. In this paper, a novel road curb detection method based on tensor voting is presented. The proposed method processes the dense point cloud acquired using a 3D LiDAR. Firstly, we utilize a sparse tensor voting approach to extract the line and surface features. Then, we use an adaptive height threshold and a surface vector to extract the point clouds of the road curbs. Finally, we utilize the height threshold to segment different obstacles from the occupancy grid map. This also provides an effective way of generating high-definition maps. The experimental results illustrate that our proposed algorithm can detect road curbs with near real-time performance.

Bohuan Xue, Jianhao Jiao, Yilong Zhu et al.

Multi-LiDAR systems have been prevalently applied in modern autonomous vehicles to render a broad view of the environments. The rapid development of 5G wireless technologies has brought a breakthrough for current cellular vehicle-to-everything (C-V2X) applications. Therefore, a novel localization and perception system in which multiple LiDARs are mounted around cities for autonomous vehicles has been proposed. However, the existing calibration methods require specific hard-to-move markers, ego-motion, or good initial values given by users. In this paper, we present a novel approach that enables automatic multi-LiDAR calibration using two poles stickered with retro-reflective tape. This method does not depend on prior environmental information, initial values of the extrinsic parameters, or movable platforms like a car. We analyze the LiDAR-pole model, verify the feasibility of the algorithm through simulation data, and present a simple method to measure the calibration errors w.r.t the ground truth. Experimental results demonstrate that our approach gains better flexibility and higher accuracy when compared with the state-of-the-art approach.

Yunfei Ye, Dong Han

Real-world data such as digital images, MRI scans and electroencephalography signals are naturally represented as matrices with structural information. Most existing classifiers aim to capture these structures by regularizing the regression matrix to be low-rank or sparse. Some other methodologies introduce factorization technique to explore nonlinear relationships of matrix data in kernel space. In this paper, we propose a multi-distance support matrix machine (MDSMM), which provides a principled way of solving matrix classification problems. The multi-distance is introduced to capture the correlation within matrix data, by means of intrinsic information in rows and columns of input data. A complex hyperplane is established upon these values to separate distinct classes. We further study the generalization bounds for i.i.d. processes and non i.i.d. process based on both SVM and SMM classifiers. For typical hypothesis classes where matrix norms are constrained, MDSMM achieves a faster learning rate than traditional classifiers. We also provide a more general approach for samples without prior knowledge. We demonstrate the merits of the proposed method by conducting exhaustive experiments on both simulation study and a number of real-word datasets.