HYHuanjing Yue

Huanjing Yue

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Huanjing Yue has a total of 109 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Test-Time Adaptation for Photovoltaic Installation Growth Assessment

    by Rui Huang, Huanjing Yue, Jianhua Guo, Xiaoqiang Lu, Jingyu Yang

    Accurately monitoring the growth of photovoltaic power station installation is crucial to assessing its socio-economic benefits and environmental impact. Researchers have constructed benchmark datasets of photovoltaic panels based on high-resolution satellite images to develop methods for detecting areas of change. However, these recent efforts confront great challenges when applied to previously unseen regions due to the limited geographical and photovoltaic panel type diversity in existing datasets, as well as the difficulty of collecting accurate pixel-level annotations. To the best of our knowledge, this work is the first to introduce test-time adaptation (TTA) to the change detection task. We propose a novel framework, termed PA-TTA (Test-Time Adaptation for Photovoltaic Installation Growth Assessment), which addresses both domain shift and catastrophic forgetting, enabling the detection of photovoltaic installation growth across diverse geographic regions without requiring ground-truth labels. Specifically, PA-TTA designs a cross-image context-aware (CCA) module based on hybrid attention, which utilizes a cross-image prototype bank to provide complementary information across images, thereby generating more accurate pseudo labels. Then, it uses the task-agnostic knowledge of the general segmentation model to adapt pseudo labels. Experimental results demonstrate that PA-TTA achieves significantly improved performance in complex and previously unseen environments. On the Telangana and Bavaria target domain data, the F1 score improves by 9.49% and 5.77%, respectively, over the second-best method. This confirms the robustness and practical applicability of the proposed approach for photovoltaic panels installation assessment without access to ground-truth labels. Code is available at https://github.com/DonMuv/PA-TTA.

  • Efficient HDR Reconstruction From Real-World Raw Images

    by Qirui Yang, Yihao Liu, Qihua Cheng, Huanjing Yue, Kun Li, Jingyu Yang

    The growing prevalence of high-resolution displays on edge devices has created a pressing need for efficient high dynamic range (HDR) imaging algorithms. However, most existing HDR methods either struggle to deliver satisfactory visual quality or incur high computational and memory costs, limiting their applicability to high-resolution inputs (typically exceeding 12 megapixels). Furthermore, current HDR dataset collection approaches are often labor-intensive and inefficient. In this work, we explore a novel and practical solution for HDR reconstruction directly from raw sensor data, aiming to enhance both performance and deployability on mobile platforms. Our key insights are threefold: (1) we propose RepUNet, a lightweight and efficient HDR network leveraging structural re-parameterization for fast and robust inference; (2) we design a new computational raw HDR data formation pipeline and construct a new raw HDR dataset, RealRaw-HDR; (3) we design a plug-and-play motion alignment loss to suppress ghosting artifacts under constrained bandwidth conditions effectively. Our model contains fewer than 830 K parameters and takes less than 3 ms to process an image of 4 K resolution using one RTX 3090 GPU. While being highly efficient, our model also achieves comparable performance to state-of-the-art HDR methods in terms of PSNR, SSIM, and a color difference metric.


  • Wavelet-Enhanced Foundation Model Adaptation for Weakly Supervised Urban Tree Canopy Mapping

    by Mingnuo Teng, Jianhua Guo, Huanjing Yue, Jingyu Yang

    Most existing weakly supervised methods for tree canopy segmentation suffer from limited accuracy due to insufficient supervision and poor generalization. We propose AMP—an end-to-end semantic segmentation framework that adaptively refines a pre-trained vision foundation model (DINOv2) via a lightweight CNN-based adapter. AMP comprises two key components: (1) a backbone that integrates high-frequency wavelet features with the foundation model through a novel High-Frequency Feature Integration (HFFI) module, and (2) a multi-branch pseudo-label generation module that produces diverse supervisory signals under sparse annotation. To mitigate the domain gap between DINOv2’s pretraining data and our remote sensing imagery, we design an adaptive adapter that fine-tunes the foundation model in a task-aware manner. Furthermore, to ensure pseudo-label reliability, we employ three parallel generation branches followed by a pixel-wise selection strategy to synthesize high-quality supervision. We have constructed a sparsely labeled dataset to enable and evaluate weakly supervised tree canopy segmentation in urban environments. Extensive results demonstrate that AMP achieves state-of-the-art performance under sparse supervision, significantly outperforming existing weakly supervised approaches. Which achieved an mIoU of 81.57% and an mF1 score of 89.43% on the UTC-sparse dataset. The code and data will be made publicly available at: https://github.com/Tengmingn/AMP.

  • DIC-DDA: Learned Asymmetric Distributed Image Compression via Dual Domain Alignment

    by Huanjing Yue, Xianghui Li, Mengyuan Sun, Gaosheng Liu, Xin Liu, Jingyu Yang

    Multi-view or stereo image compression is an essential technology in 3D related applications. Due to the overlap between different views, exploring their correlations can help improve the compression rate. However, the computing complexity of joint encoding at the encoding side is a heavy burden for terminal encoders. To solve this problem, the learned Distributed Image Coding (DIC), which only uses the correlated view (namely the side image, SI) in the decoder side, has gained much attention in recent years. In this work, we explore asymmetric DIC where one view is selected as the SI and is losslessly compressed. The key problem in learned asymmetric DIC is alignment between the transmitted low-quality target image and high-quality SI. Previous methods usually adopt patch-level alignment with the offset index obtained from degraded (via re-encoded and decoded) SI and the decoded target image, which hinders the alignment accuracy. In this work, we propose a dual domain alignment strategy, which includes degraded domain and fused domain pixel-wise offset estimation. For the degraded domain alignment, we estimate the offset between the degraded SI feature and the degraded target image feature, which eliminates the difficulties in cross-domain matching. For the fused-domain alignment, we observe that the fusion result of degraded target feature and aligned side image feature implicitly contains fine-scale disparity information. Therefore, we estimate the fine-scale offset from the fusion result, which helps refine the degraded domain offsets. We further propose a selective enhancement module to repair the mismatched region in the aligned feature. Extensive experiments on three datasets demonstrate the superiority of our proposed method, outperforming the second-best method by 16% in terms of average BD-rate reduction on the KITTI Stereo dataset. Our code will be released after the acceptance of this work.

  • Open Access

    Zero-Shot Video Restoration and Enhancement with Assistance of Video Diffusion Models

    by Cong Cao, Huanjing Yue, Shangbin Xie, Xin Liu, Jingyu Yang

    Although diffusion-based zero-shot image restoration and enhancement methods have achieved great success, applying them to video restoration or enhancement will lead to severe temporal flickering. In this paper, we propose the first framework that utilizes the rapidly-developed video diffusion model to assist the image-based method in maintaining more temporal consistency for zero-shot video restoration and enhancement. We propose homologous latents fusion, heterogenous latents fusion, and a COT-based fusion ratio strategy to utilize both homologous and heterogenous text-to-video diffusion models to complement the image method. Moreover, we propose temporal-strengthening post-processing to utilize the image-to-video diffusion model to further improve temporal consistency. Our method is training-free and can be applied to any diffusion-based image restoration and enhancement methods. Experimental results demonstrate the superiority of the proposed method.

  • TransDiff: Unsupervised Non-Line-of-Sight Imaging With Aperture-Limited Relay Surfaces

    by Xingyu Cui, Huanjing Yue, Shida Sun, Yue Li, Yusen Hou, Zhiwei Xiong, Jingyu Yang

    Non-line-of-sight (NLOS) imaging aims to reconstruct scenes hidden from direct view and has broad applications in robotic vision, rescue operations, autonomous driving, and remote sensing. However, most existing methods rely on densely sampled transients from large, continuous relay surfaces, which limits their practicality in real-world scenarios with aperture constraints. To address this limitation, we propose an unsupervised zero-shot framework tailored for confocal NLOS imaging with aperture-limited relay surfaces. Our method leverages latent diffusion models to recover fully-sampled transients from undersampled versions by enforcing measurement consistency during the sampling process. To further improve recovered transient quality, we introduce a progressive recovery strategy that incrementally recovers missing transient values, effectively mitigating the impact of severe aperture limitations. In addition, to suppress error propagation during recovery, we develop a backpropagation-based error correction reconstruction algorithm that refines intermediate recovered transients by enforcing sparsity regularization in the voxel domain, enabling high-fidelity final reconstructions. Extensive experiments on both simulated and real-world datasets validate the robustness and generalization capability of our method across diverse aperture-limited relay surfaces. Notably, our method follows a zero-shot paradigm, requiring only a single pretraining stage without paired data or pattern-specific retraining, making it a more practical and generalizable framework for NLOS imaging.


  • Open Access

    DSDNet: Raw Domain Demoiréing via Dual Color-Space Synergy

    by Qirui Yang, Fangpu Zhang, Yeying Jin, Qihua Cheng, Pengtao Jiang, Huanjing Yue, Jingyu Yang


  • Open Access

    FEALLM: Advancing Facial Emotion Analysis in Multimodal Large Language Models with Emotional Synergy and Reasoning

    by Zhuozhao Hu, Kaishen Yuan, Xin Liu, Zitong Yu, Yuan Zong, Jingang Shi, Huanjing Yue, Jingyu Yang


  • Reprojection-Guided Non-Line-of-Sight Imaging Under Irregular Undersampling

    by Xingyu Cui, Huanjing Yue, Jingyu Yang

    Non-line-of-sight (NLOS) imaging enables the visualization of scenes outside the direct line of sight, with broad potential applications in robotic vision, autonomous navigation, disaster relief, and medical diagnostics. However, most existing NLOS methods depend on densely sampled transients collected from large and continuous relay surfaces, which restricts their applicability in real-world scenarios where the available relay surfaces are typically sparse, irregular, or fragmented. To overcome the challenges posed by irregular undersampling in NLOS imaging, we propose a reprojection-guided framework designed for robust performance in such scenarios. At its core is a transient recovery network that reconstructs denoised, fully-sampled transients from noisy, undersampled inputs, thereby mitigating the inherent ill-posedness of the problem. To further enhance recovery quality and improve generalization, we develop a range-space reprojection (RSRP) module that extracts intrinsic information from the measurements and generates explicit guidance. This guidance is leveraged by the spatio-temporal modulation (STM) blocks, which adaptively control its influence to enhance transient denoising and recovery. Extensive experiments on both simulated and real-world datasets demonstrate that our method significantly outperforms existing approaches, showing strong generalization across diverse relay surfaces and marking a substantial step toward practical and robust NLOS imaging.