Yaocong Hu

Yaocong Hu

Auto-generated profile, based on publicly available information

Yaocong Hu

Key Laboratory of Advanced Perception and Intelligent Control of High-end Equipment, Ministry of Education, Anhui Polytechnic University, Wuhu, China

Research Focus & Metrics

Academic Career Length

4Years (2023-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 March 2026 in IEEE Transactions on Circuits and Systems for Video Technology
Verified
Journal Article
Closed access
YHYaocong Hu
Yaocong Hu
MJMengbo Jia
Mengbo Jia
PWPindeng Wang
Pindeng Wang
NJNing Jiang
Ning Jiang
KXKe Xu
Ke Xu
Teng LiTeng Li
Teng Li
Crowd counting has long been a crucial topic in the domains of computer vision and video surveillance. In particular, with the widespread use of thermal or depth cameras, RGB-X crowd counting has emerged as a prominent research focus. Although depth or infrared images provide complementary information, the core challenge remains in effectively unifying heterogeneous cross-modality and cross-scale information to form a comprehensive representation of crowd distributions in complex scenes. To address this problem, we propose a novel Mamba-based framework, termed UMS-VMamba-CC for multi-modal (RGB-X) crowd counting. Specifically, we design the cross-modality disentanglement fusion visual Mamba (CMDF-VMamba) that uses self-supervised learning to decompose modality-invariant and modality-specific features in spatial-frequency domains, followed by multi-modal feature aggregation via the gating mechanism. For cross-scale fusion, we design the cross-scale pyramid fusion visual Mamba (CSPF-VMamba), which adopts the bi-directional pyramid structure to incorporate low-level features into high-level representations during downsampling, while upsampling high-level features and aggregating them with low-level features through state space contextual modeling. Comprehensive experiments on multiple mainstream datasets demonstrate that the UMS-VMamba-CC framework achieves competitive performance for RGB-X crowd counting.

24 February 2026 in IEEE Sensors Journal
Verified
Journal Article
Closed access
ZXZile Xiao
Zile Xiao
YZYuhan Zhao
Yuhan Zhao
CNChuanzhi Ning
Chuanzhi Ning
YHYaocong Hu
Yaocong Hu
JTJing Tian
Jing Tian
BLBingyou Liu
Bingyou Liu
RGB-thermal (RGB-T) tracking leverages the complementary advantages of visible light and thermal infrared sensors to enhance robustness under varying illumination, occlusion, and background clutter. However, existing fusion methods still suffer from redundant feature representations and insufficient cross-modal interaction, which constrain their adaptability in complex environments. To overcome these challenges, we propose an Adaptive Cross-Modality Interaction (ACMI) framework based on the vision transformer (ViT), which enables dynamic and bidirectional information exchange between red–green–blue (RGB) and thermal infrared( TIR) modalities. Within ACMI, Shallow Feature Enhancement (SFE) submodule adaptively reweights modality-specific shallow features in a data-driven manner to strengthen discriminative representations; the DynamicFusion module jointly models spatial–channel dependencies to enhance informative cross-sensor correlations; SEAdapter introduces residual calibration to adaptively balance modality contributions and mitigate information loss. These modules are embedded throughout the Transformer layers, forming a hierarchical architecture that progressively strengthens cross-modal feature understanding. Extensive experiments demonstrate that ACMI achieves consistent performance improvements across multiple RGB-T tracking benchmarks, particularly in low-resolution and cluttered scenes, while maintaining real-time performance. The results validate ACMI as a robust and adaptive framework for efficient multi-sensor visual tracking.

20 October 2023
Verified
Conference Paper
Open access
YLYuanyuan Lin
Yuanyuan Lin
HYHuicheng Yang
Huicheng Yang
YHYaocong Hu
Yaocong Hu
ZSZhen Shuai
Zhen Shuai
WLWenting Li
Wenting Li
Crowd counting apply in many areas. In order to prevent stampede accidents, it is necessary to accurately forecast the number of individuals in some scenarios with large numbers. Among them, CNN and Transformer are the mainstream models to solve crowd counting in dense scenes. CNNs are good at capturing local features, but they are weak at capturing global features. Transformer uses self-attention module to extract global features, but the model will lead to the loss of local details. Therefore, both CNN models and transformer models have certain limitations in the task of estimating people in dense scenes. hence this text presents a method combining CNN and Transformer to forecast the number of individuals in dense scenarios. First, the density plot is regressed at the CNN layer, to determine the number of individuals, you can sum the values from the density plot. The headcount is directly generated at the Transformer layer. Secondly, a weight term is proposed to fuse the number of people generated by the CNN layer and the Transformer layer. In the departed methods of predicting the headcount, most of them were summed into density plots to get the headcount, but we have made an innovation in the Transformer counting branch to directly generate the headcount at the Transformer layer, reducing the number of parameters and speeding up training. Experiments show that the proposed method has been found to achieve satisfactory performance on the dataset.

20 October 2023
Verified
Conference Paper
Open access
WLWenting Li
Wenting Li
HYHuicheng Yang
Huicheng Yang
YHYaocong Hu
Yaocong Hu
YLYuanyuan Lin
Yuanyuan Lin
ZSZhen Shuai
Zhen Shuai
Aiming at the problem that it is difficult to balance the accuracy and computational efficiency of the road scene semantic segmentation algorithm, a real-time road scene segmentation algorithm based on knowledge distillation is proposed. The algorithm introduces three key functional modules; (1) using an asymmetric convolutional pyramid module on the top of the encoder; (2) incorporating a coordinate attention module in the network; (3) Using high accuracy non-real time models to distill knowledge on lightweight model. The method in this paper achieves a balance of accuracy and real-time performance on the CamVid dataset. the inference speed is 63.7 FPS and the GFLOPs are 109.2 while the mIoU is 79.3% when computed on an NVIDIA GTX 1080Ti.

9 April 2026 in IEEE Transactions on Intelligent Transportation Systems
Verified
Journal Article
Closed access
YHYaocong Hu
Yaocong Hu
PWPindeng Wang
Pindeng Wang
MJMengbo Jia
Mengbo Jia
YZYifan Zhang
Yifan Zhang
JHJinwen Hong
Jinwen Hong
Teng LiTeng Li
Teng Li
Road scene semantic segmentation is one of the most crucial tasks in autonomous driving. Unlike general semantic segmentation, traffic applications require fast image processing, making it difficult to satisfy both high-performance and real-time. Moreover, how to efficiently fuse multi-scale feature representations for a comprehensive understanding of the road scene remains challenging. To address these issues, we propose a lightweight U-shape deep learning framework for high-performance and real-time road scene semantic segmentation in this paper. Specifically, we design the lightweight mixed convolutional block (LMCB), which combines depthwise separable convolution and asymmetric dilated convolution, with the purpose of reducing computational cost and expanding the receptive field for road scene feature extraction. Then, efficient multi-scale feature fusion Transformer (EMFFTrans) is proposed to bridge the encoder-decoder interaction, leveraging the separable cross-channel attention (SCCA) mechanism to capture the inherent dependencies across multi-scale feature maps, thereby facilitating comprehensive representation for road scenes. Experimental results show that our proposed framework, running on a single RTX 3090 GPU card, strikes a satisfactory balance between accuracy and efficiency. Specifically, the LMCB-EMFFTrans-Slim model achieves 73.7% mIOU at 166 FPS on the CamVid dataset and 76.3% mIOU at 113 FPS on the Cityscapes dataset with fewer than 1 million learnable parameters. The code will be soon available at: https://github.com/WH-IPUSLab/EMFFTrans