YSYang Sun

Yang Sun

*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


Yang Sun has a total of 18 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Controlled Evaluation of Dual-Channel Feature Enhancement and Multi-Level Knowledge Distillation for Lightweight Plant Disease Recognition

    by Xin Lei, Yonghuai Liu, Ardhendu Behera, Reena Reena, Yang Sun, Fuzhong Li, Wuping Zhang, Chao Lei

    Plant disease symptoms combine local texture changes with patterns distributed across a leaf, while practical recognition models must remain compact. We introduce DC-FEN, a MobileNetV3-based design that models spatial-token relations and channel interactions in parallel and injects them through gated residual fusion. We also examine output-distribution, direct-feature, and token-relation transfer under same-backbone and heterogeneous teachers. PlantVillage and Plant Pathology 2021 (FGVC8) are evaluated with duplicate-audited, group-aware 70/15/15 splits, an explicit unresolved-leaf sensitivity check, validation-only selection, five training seeds, class-sensitive metrics, and paired seed-wise descriptive summaries. On PlantVillage, the no-additional-attention student, DC-FEN teacher, and DC-FEN joint student obtain macro F1 scores of 96.46±0.91%, 96.90±0.40%, and 96.55±0.25%. On FGVC8, the corresponding scores are 87.29±0.63%, 87.14±0.52%, and 87.20±0.26%. At the prespecified FGVC8 threshold of 0.5, DCAB changed sample-wise F1 by −0.02±0.55 percentage points relative to the unmodified backbone; validation-selected global and label-specific thresholds changed this contrast to +0.28±0.55 and +0.55±0.29 points, while threshold-free macro mAP remained essentially unchanged. A duplicate-audited PlantDoc pressure test reduced frozen-checkpoint accuracy to 30.34±1.10% and 29.57±1.10%, showing that external generalization remains unestablished. A ResNet50 teacher gives logit-only students 97.42±0.51% macro F1 on PlantVillage and 89.82±0.43% sample-wise F1 on FGVC8. After separately weighting the direct and relation terms, the corresponding joint students obtain 97.37±0.56% and 89.94±0.27%, recovering the degradation seen with unit internal weights while remaining close to logit-only transfer. Thus, the study evaluates the benefits and limits of explicit spatial–channel interaction and shows that adding intermediate transfer constraints does not guarantee a stronger student.

  • VIM-Net: A voxel-interaction multimodal network for 3D object detection

    by Minghan Wang, Bin Lu, Yang Sun, Xijiong Wang, Yonghuai Liu, Ardhendu Behera, Baowen Zhang

  • Open Access

    Dual-Channel Attention and Category-Guided Distillation for Lightweight Plant Disease Recognition

    by Xin Lei, Yonghuai Liu, Ardhendu Behera, Reena Reena, Yang Sun, Fuzhong Li, Wuping Zhang, Chao Lei

    Plant disease recognition remains challenging due to complex imaging conditions, inter-class visual similarity, and the need to deploy accurate models on resource-constrained devices. Although convolutional neural networks (CNNs) are effective at local pattern extraction, their locality bias limits global context modeling. Transformer-based models alleviate this issue, yet many existing designs do not explicitly exploit the complementarity between spatial and channel interactions. To address this limitation, we propose a Dual-Channel Feature Enhancement Network (DC-FEN) as the teacher model, which integrates spatial attention and channel attention to capture lesion-sensitive spatial cues and discriminative channel semantics. To enable lightweight deployment, we further introduce a Category-Guided Knowledge Distillation (CGKD) framework that transfers both category-level semantic knowledge and feature-level relational knowledge from the teacher to a MobileNetV3 student. Experimental results show that the proposed distillation framework consistently improves the performance of the lightweight MobileNetV3 student over its original baseline.

  • Exploiting Label Uncertainty for Enhanced 3D Object Detection From Point Clouds

    by Yang Sun, Bin Lu, Yonghuai Liu, Zhenyu Yang, Ardhendu Behera, Ran Song, Hejin Yuan, Haiyan Jiang

    Accurate detection of objects from LiDAR point clouds is crucial for autonomous driving and environment modeling. However, uncertainties in ground truth labels due to occlusions, sparsity, and truncation can hinder model training and performance. This paper introduces two strategies to address these issues: 1) Soft Regression Loss (SoRL) and 2) Discrete Quantization Sampling (DQS). SoRL utilizes Gaussian distributions for object predictions, measuring uncertainty based on the probability of ground truth labels within these distributions. This method effectively accounts for deviations in object location and orientation. Meanwhile, DQS introduces uncertainty scores for dynamic sample selection, aiming to refine the quality of positive samples for regression. Based on the proposed modules, we design a lightweight multi-stage object detection framework. Notably, these modules can enhance existing 3D object detection methods without affecting significantly inference speeds. Experiments over benchmark datasets show the effectiveness of our method, especially for cars in sparse point clouds.


  • Voxel Graph Attention for 3-D Object Detection From Point Clouds

    by Bin Lu, Yang Sun, Zhenyu Yang

    Recently, 3D object detection from LiDAR point clouds has advanced rapidly, but the detection performance still has shortcomings in complex scenarios. In this paper, we introduce a novel two-stage 3D object detector that utilizes only raw point clouds as input. Specifically, we propose two novel strategies that can effectively enhance the detection performance: the Feature Interpolation Aggregation (FIA) and Graph Attention Pooling (GAP). The former is used to initialize the graph centroid features across different resolutions, which is essential for constructing local graphs. The latter constructs hierarchical local graphs on multi-scale voxels to extract region of interest (RoI) features by considering the contributions of neighboring voxels. This enables the aggregation of more effective and robust features for detection. The proposed innovations enhance the detector’s feature extraction capability and improve its robustness in complex scenarios. Experimental results on the KITTI and Waymo datasets demonstrate that VoxelGraph-RCNN achieves competitive performance compared with state-of-the-art detectors. Specifically, the proposed model surpasses the baseline Voxel R-CNN by 0.43%, 1.04%, and 0.87% in detecting objects with easy, moderate, and hard difficulty levels on the KITTI datasets, respectively.