HLHao Li

Hao Li

*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


Hao Li has a total of 65 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    The 1st International Workshop on Agentic Multimodal Intelligence: Models, Benchmarks, and Applications

    by Qi Chen, Junyan Wang, Yuankai Qi, Jiajun Deng, Hao Li, Mingkui Tan, Houqiang Li, Xiaojun Chang, Qi Wu, Anton van Den Hengel

  • EPro-PnP: Generalized End-to-End Probabilistic Perspective-n-Points for Monocular Object Pose Estimation

    by Hansheng Chen, Wei Tian, Pichao Wang, Fan Wang, Lu Xiong, Hao Li

    Locating 3D objects from a single RGB image via Perspective-n-Point (PnP) is a long-standing problem in computer vision. Driven by end-to-end deep learning, recent studies suggest interpreting PnP as a differentiable layer, allowing for partial learning of 2D-3D point correspondences by backpropagating the gradients of pose loss. Yet, learning the entire correspondences from scratch is highly challenging, particularly for ambiguous pose solutions, where the globally optimal pose is theoretically non-differentiable w.r.t. the points. In this paper, we propose the EPro-PnP, a probabilistic PnP layer for general end-to-end pose estimation, which outputs a distribution of pose with differentiable probability density on the SE(3) manifold. The 2D-3D coordinates and corresponding weights are treated as intermediate variables learned by minimizing the KL divergence between the predicted and target pose distribution. The underlying principle generalizes previous approaches, and resembles the attention mechanism. EPro-PnP can enhance existing correspondence networks, closing the gap between PnP-based method and the task-specific leaders on the LineMOD 6DoF pose estimation benchmark. Furthermore, EPro-PnP helps to explore new possibilities of network design, as we demonstrate a novel deformable correspondence network with the state-of-the-art pose accuracy on the nuScenes 3D object detection benchmark.


  • Open Access

    EPro-PnP: Generalized End-to-End Probabilistic Perspective-n-Points for Monocular Object Pose Estimation

    by Hansheng Chen, Wei Tian, Pichao Wang, Fan Wang, Lu Xiong, Hao Li

    Locating 3D objects from a single RGB image via Perspective-n-Point (PnP) is a long-standing problem in computer vision. Driven by end-to-end deep learning, recent studies suggest interpreting PnP as a differentiable layer, allowing for partial learning of 2D-3D point correspondences by backpropagating the gradients of pose loss. Yet, learning the entire correspondences from scratch is highly challenging, particularly for ambiguous pose solutions, where the globally optimal pose is theoretically non-differentiable w.r.t. the points. In this paper, we propose the EPro-PnP, a probabilistic PnP layer for general end-to-end pose estimation, which outputs a distribution of pose with differentiable probability density on the SE(3) manifold. The 2D-3D coordinates and corresponding weights are treated as intermediate variables learned by minimizing the KL divergence between the predicted and target pose distribution. The underlying principle generalizes previous approaches, and resembles the attention mechanism. EPro-PnP can enhance existing correspondence networks, closing the gap between PnP-based method and the task-specific leaders on the LineMOD 6DoF pose estimation benchmark. Furthermore, EPro-PnP helps to explore new possibilities of network design, as we demonstrate a novel deformable correspondence network with the state-of-the-art pose accuracy on the nuScenes 3D object detection benchmark. Our code is available at https://github.com/tjiiv-cprg/EPro-PnP-v2.

  • Improved DeepSORT-based pedestrian tracking and recognition method with fused face information

    by Hao Li, Zhijun Chen, Yu Zhang, Dayang Li, Kun Yang

    In order to solve the problem of continuity and accuracy of regional pedestrian identification, a method for regional identification and tracking of specific pedestrians with fusion of face information is proposed by using the front and rear frame information of the video stream. The FaceNet face identity network is constructed using the lightweight MobilenetV3-Large feature extraction network, and a fully connected layer is added for classification. Cross-Entropy Loss and Triplet Loss are integrated to build a classifier for stable loss convergence. The identity recognition method is introduced into DeepSORT, and an identity constraint module is added to solve the IDswitch caused by changes in appearance or pedestrian direction. This method combines the two modules of identification network and pedestrian tracking, and then realizes the continuous identification and tracking of pedestrians in the area. The experimental results show that: in terms of identity recognition, the accuracy of the improved system is 11% higher than that of the original algorithm in the autonomous test link; the accuracy of the tracking model is 69.S3% in MOTA on the MOT16 dataset, and the improved system is compared with the original algorithm in the autonomous test. During the test, MOTA increased from 72.38% to 89.84%, while IDswitch was greatly reduced. It realizes the detection and continuous tracking of specific pedestrians in the area, and provides technical support for the behavior of pedestrians in the area.

  • What Limits the Performance of Local Self-attention?

    by Jingkai Zhou, Pichao Wang, Jiasheng Tang, Fan Wang, Qiong Liu, Hao Li, Rong Jin


  • Open Access

    MimCo: Masked Image Modeling Pre-training with Contrastive Teacher

    by Qiang Zhou, Chaohui Yu, Hao Luo, Zhibin Wang, Hao Li

    Recent masked image modeling (MIM) has received much attention in self-supervised learning (SSL), which requires the target model to recover the masked part of the input image. Although MIM-based pre-training methods achieve new state-of-the-art performance when transferred to many downstream tasks, the visualizations show that the learned representations are less separable, especially compared to those based on contrastive learning pre-training. This inspires us to think whether the linear separability of MIM pre-trained representation can be further improved, thereby improving the pre-training performance. Since MIM and contrastive learning tend to utilize different data augmentations and training strategies, combining these two pretext tasks is not trivial. In this work, we propose a novel and flexible pre-training framework, named MimCo, which combines MIM and contrastive learning through two-stage pre-training. Specifically, MimCo takes a pre-trained contrastive learning model as the teacher model and is pre-trained with two types of learning targets: patch-level and image-level reconstruction losses. Extensive transfer experiments on downstream tasks demonstrate the superior performance of our MimCo pre-training framework. Taking ViT-S as an example, when using the pre-trained MoCov3-ViT-S as the teacher model, MimCo only needs 100 epochs of pre-training to achieve 82.53% top-1 finetuning accuracy on Imagenet-1K, which outperforms the state-of-the-art self-supervised learning counterparts.


  • Contrastive Haze-Aware Learning for Dynamic Remote Sensing Image Dehazing

    by Jiangtao Nie, Wei Wei, Lei Zhang, Jianlong Yuan, Zhibin Wang, Hao Li

    Image dehazing methods aim to recover a clear image from its hazy counterpart. While various dehazing methods have been proposed, their performance on real-world remote sensing (RS) images remains unsatisfying. A key reason is that the complex weather and imaging conditions (e.g., large fields of view) cause the haze condition to dramatically change in different images, while most existing methods fail to flexibly adapt their dehazing model to the specific haze condition in each image. To mitigate this problem, we present a contrastive haze-aware learning based dynamic dehazing method which demonstrates two aspects of advantage. On one hand, a contrastive clustering scheme is utilized to learn the image-wise haze representation using a set of real-world hazy images in an unsupervised manner, which enables identifying and discriminating the specific haze condition in each given hazy image. On the other hand, with the learned haze representation, a parameter generator can produce haze-aware parameters to dynamically construct a dehazing model for the given hazy image, which empowers us to adaptively dehaze the image based on its specific haze condition and thus improves the generalization ability. In addition, a new contrastive loss defined based on the learned haze representation is further utilized for model training and leads to better performance. To demonstrate the effectiveness of the proposed method, we evaluate it on two benchmark RS image datasets including various real-world hazy images, and observe obviously superiority over other state-of-the-art competitors.


  • KVT: k-NN Attention for Boosting Vision Transformers

    by Pichao Wang, Xue Wang, Fan Wang, Ming Lin, Shuning Chang, Hao Li, Rong Jin


  • Unstructured Feature Decoupling for Vehicle Re-identification

    by Wen Qian, Hao Luo, Silong Peng, Fan Wang, Chen Chen, Hao Li


  • TransFGU: A Top-Down Approach to Fine-Grained Unsupervised Semantic Segmentation

    by Zhaoyuan Yin, Pichao Wang, Fan Wang, Xianzhe Xu, Hanling Zhang, Hao Li, Rong Jin