Chen Xie

Mr. Chen Xie

orcid

Research Focus & Metrics

Academic Career Length

1Year (2026-Present)

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

Representative Publications

View All Publications
19 June 2024 in IEEE Access
Verified
Journal Article
Open access
HPHongming Peng
Hongming Peng
SXSiyu Xiang
Siyu Xiang
MCMingju Chen
Mingju Chen
HLHongyang Li
Hongyang Li
QSQin Su
Qin Su
Semantic segmentation is an important task in the field of computer vision. However, Deeplabv3+, which is commonly used for semantic segmentation, has shortcomings, such as a large number of parameters and a tendency to lose detailed information. Therefore, this paper proposes DCN-Deeplabv3+, an improved road segmentation algorithm with dual attention modules based on the Deeplabv3+ network. (1) MobileNetV2 is used as the backbone network to reduce model parameters and memory consumption. (2) DenseASPP+SP is used for multi-scale information fusion to obtain a larger sensory field for improved model performance. (3) The deep learning model’s understanding of the spatial structure of the input data is enhanced by using CA (coordinate attention) to improve the model’s performance in dealing with spatial structure-related tasks. (4) The neural attention mechanism (NAM) is applied to better focus on key regions in the image, thereby improving the accuracy of target detection. The experimental results show that mIoU and mPA are improved by 1.20% and 2.30% on the PASCAL VOC 2012 dataset, mIoU and mPA are improved by 3.15% and 3.90% on the Cityscapes dataset, respectively. It can be concluded that the method proposed in this paper outperforms the baseline method and exhibits excellent semantic segmentation precision.

16 October 2024 in Electronics
Verified
Journal Article
Open access
PDF file
SPShangzhen Pang
Shangzhen Pang
TTTzer Hwai Gilbert Thio
Tzer Hwai Gilbert Thio
Fei Lu SiawFei Lu Siaw
Fei Lu Siaw
MCMingju Chen
Mingju Chen
XYXIA YULE
XIA YULE
The classic interactive image segmentation algorithm GrabCut achieves segmentation through iterative optimization. However, GrabCut requires multiple iterations, resulting in slower performance. Moreover, relying solely on a rectangular bounding box can sometimes lead to inaccuracies, especially when dealing with complex shapes or intricate object boundaries. To address these issues in GrabCut, an improvement is introduced by incorporating appearance overlap terms to optimize segmentation energy function, thereby achieving optimal segmentation results in a single iteration. This enhancement significantly reduces computational costs while improving the overall segmentation speed without compromising accuracy. Additionally, users can directly provide seed points on the image to more accurately indicate foreground and background regions, rather than relying solely on a bounding box. This interactive approach not only enhances the algorithm’s ability to accurately segment complex objects but also simplifies the user experience. We evaluate the experimental results through qualitative and quantitative analysis. In qualitative analysis, improvements in segmentation accuracy are visibly demonstrated through segmented images and residual segmentation results. In quantitative analysis, the improved algorithm outperforms GrabCut and min_cut algorithms in processing speed. When dealing with scenes where complex objects or foreground objects are very similar to the background, the improved algorithm will display more stable segmentation results.

24 May 2023 in Applied Sciences
Verified
Journal Article
Open access
PDF file
MCMingju Chen
Mingju Chen
ZDZHENGXU DUAN
ZHENGXU DUAN
ZLZhongxiao Lan
Zhongxiao Lan
SYSIHANG YI
SIHANG YI
At present, Chinese 3D reconstruction solutions using stereo cameras mainly face known, indoor, structured scenes; for the reconstruction of unstructured, larger-scale scenes with a large variety of texture information of different intensities, there are certain difficulties in ensuring accuracy and real-time processing. For the above problems, we propose a scene reconstruction method using stereo vision. Firstly, considering the influence of outdoor lighting and weather on the captured 2D images, the optimized SAD-FAST feature detection algorithm and stereo-matching strategy were employed in the stereo-matching stage to improve the overall efficiency and matching quality at this stage. Then, a homogenized feature extraction algorithm with gradient value decreasing step by step (GVDS) was used in the depth value calculation to ensure a sufficient number of feature points for strong texture information while extracting features from weak-texture areas, which greatly improved the quality and speed of unstructured scene reconstruction. We conducted experiments to validate the proposed method, and the results showed the feasibility of the proposed method and its high practical value.

2 August 2024 in IEEE Access
Verified
Journal Article
Open access
QSQin Su
Qin Su
JZJuling Zhang
Juling Zhang
MCMingju Chen
Mingju Chen
HPHongming Peng
Hongming Peng
To address the detection challenges of keypoints, such as misdetections and omissions caused by backgrounds, occlusions, small targets, and extreme viewpoints in complex electrical power operation environments for power workers. This study proposes a 2D pose estimation algorithm for power workers based on YOLOv5s6-Pose: PW-YOLO-Pose. In this study, the detection rate of occluded keypoints is improved by embedding the Swin Transformer encoder in the top layer of the backbone network. The proposed BiFPN (a weighted bi-directional feature pyramid network) structure with a small target detection layer improves the detection rate of small target characters and the precision of their keypoints’detection. The keypoint regression precision is improved overall by using CA (coordinate attention) in the model neck and improving the bounding box regression loss function to Wise-IoU. The algorithmic model in this study demonstrates excellent detection and largely meets the real-time requirements on the proposed power worker pose estimation dataset in this study. The mAP 0.5 (The mean average precision when the threshold for object keypoint similarity is set to 0.5.)and mAP 0.5:0.95 are 93.35% and 64.75% respectively, which are 5.22% and 1.53% higher than the baseline model. The detection time of a single image is 21.3 ms, respectively. It can serve as a valuable theoretical foundation and reference for behavior recognition and state monitoring of power workers in intricate electrical power operation environments.

8 June 2022 in Journal of Circuits, Systems and Computers
Verified
Journal Article
Closed access
PDF file
ZCZhengwei Chang
Zhengwei Chang
YDYuanshi Deng
Yuanshi Deng
JWJie Wu
Jie Wu
XXXingzhong Xiong
Xingzhong Xiong
MCMingju Chen
Mingju Chen
XXXiaona Xie
Xiaona Xie

18 August 2023 in Sensors
Verified
Journal Article
Open access
PDF file
MCMingju Chen
Mingju Chen
SYSIHANG YI
SIHANG YI
ZLZhongxiao Lan
Zhongxiao Lan
ZDZHENGXU DUAN
ZHENGXU DUAN
Blurring is one of the main degradation factors in image degradation, so image deblurring is of great interest as a fundamental problem in low-level computer vision. Because of the limited receptive field, traditional CNNs lack global fuzzy region modeling, and do not make full use of rich context information between features. Recently, a transformer-based neural network structure has performed well in natural language tasks, inspiring rapid development in the field of defuzzification. Therefore, in this paper, a hybrid architecture based on CNN and transformers is used for image deblurring. Specifically, we first extract the shallow features of the blurred images using a cross-layer feature fusion block that emphasizes the contextual information of each feature extraction layer. Secondly, an efficient transformer module for extracting deep features is designed, which fully aggregates feature information at medium and long distances using vertical and horizontal intra- and inter-strip attention layers, and a dual gating mechanism is used as a feedforward neural network, which effectively reduces redundant features. Finally, the cross-layer feature fusion block is used to complement the feature information to obtain the deblurred image. Extensive experimental results on publicly available benchmark datasets GoPro, HIDE, and the real dataset RealBlur show that the proposed method outperforms the current mainstream deblurring algorithms and recovers the edge contours and texture details of the images more clearly.