Junan Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Junan Chen has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
- Dandan Fan
School of Data Science, The Chinese University of Hong Kong, Shenzhen, Guangdong, China
2 shared publications - David Zhang
The Chinese University of Hong Kong
2 shared publications - Xu Liang
Harbin Institute of Technology
2 shared publications - Baoyuan Wu
School of Data Science, The Chinese University of Hong Kong, Shenzhen, China
1 shared publication - Bob Zhang
University of Macau, Macau SAR, China
1 shared publication - Chunsheng Zhang
School of Data Science, The Chinese University of Hong Kong, Shenzhen (CUHK-Shenzhen), Shenzhen, China
1 shared publication
Latest Publications
- 15 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-13
AMCOA: Agent Motion Control With Obstacle Avoidance for Efficient Palm ROI Extraction
Palm recognition facilitates various real-world applications such as security access control and identity verification. The contactless method of acquiring palm images significantly reduces the risk of bacterial transmission and enhances user-friendliness. However, this approach also presents inherent difficulties and challenges. Accurately identifying the region of interest (ROI) is crucial for feature extraction and recognition. The absence of strict guidelines regarding the position and angle of the user’s palm during contactless image acquisition complicates the accurate localization and extraction of the ROI. To address this issue, we propose an innovative ROI extraction method that utilizes Agent Motion Control with Obstacle Avoidance (AMCOA). The agent dynamically identifies finger valley points while navigating and avoiding obstacles in the image, enabling the effective extraction of both the palmprint and palm vein ROIs based on the positions of two keypoints within the finger valleys. Our method is implemented in a synchronous acquisition system designed for palmprint and palm vein ROIs. We use a sensor module that simultaneously captures the distance, infrared (IR) image, and visible light (VL) image of the hand. The IR image is used to segment the hand area and extract the palm veins ROI, while the VL image aids in the extraction of the palmprint ROI after proper alignment. Experimental results indicate that our proposed method achieves both accurate and efficient ROI extraction.
- 15 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 74 (00189456) , 1-13
AMCOA: Agent Motion Control With Obstacle Avoidance for Efficient Palm ROI Extraction
Palm recognition facilitates various real-world applications such as security access control and identity verification. The contactless method of acquiring palm images significantly reduces the risk of bacterial transmission and enhances user-friendliness. However, this approach also presents inherent difficulties and challenges. Accurately identifying the region of interest (ROI) is crucial for feature extraction and recognition. The absence of strict guidelines regarding the position and angle of the user’s palm during contactless image acquisition complicates the accurate localization and extraction of the ROI. To address this issue, we propose an innovative ROI extraction method that utilizes Agent Motion Control with Obstacle Avoidance (AMCOA). The agent dynamically identifies finger valley points while navigating and avoiding obstacles in the image, enabling the effective extraction of both the palmprint and palm vein ROIs based on the positions of two keypoints within the finger valleys. Our method is implemented in a synchronous acquisition system designed for palmprint and palm vein ROIs. We use a sensor module that simultaneously captures the distance, infrared (IR) image, and visible light (VL) image of the hand. The IR image is used to segment the hand area and extract the palm veins ROI, while the VL image aids in the extraction of the palmprint ROI after proper alignment. Experimental results indicate that our proposed method achieves both accurate and efficient ROI extraction.
- 23 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-13
HGAIQA: A Novel Hand-Geometry-Aware Image Quality Assessment Framework for Contactless Palmprint Recognition
Contactless palmprint recognition has gained traction due to its convenience and hygienic benefits. However, in real-world scenarios with complex backgrounds and varying hand poses, evaluating image quality to enhance recognition performance remains a significant challenge. To address this, we propose a novel Hand-Geometry Aware Contactless Palmprint Image Quality Assessment (HGAIQA) framework. Unlike existing methods that assess only the palmprint ROI, our framework evaluates the entire image. Firstly, it employs a high-resolution hand segmentation network and keypoint heatmap module to identify hand region and joint keypoints. Secondly, it evaluates the palm’s flatness based on geometric features and assesses additional quality attributes such as brightness and sharpness. Lastly, it determines image quality by analyzing the intra-class and inter-class distributions of fused multi-features. After integrating with subsequent ROI localization and recognition algorithms, experiments show a substantial 21.2% reduction in EER for palmprint recognition on the COEP database by removing the lowest 10% of low-quality images. These results demonstrate the effectiveness of our approach in significantly enhancing palmprint recognition performance.
- 23 October 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Instrumentation and Measurement
- Vol. 73 (00189456) , 1-13
HGAIQA: A Novel Hand-Geometry-Aware Image Quality Assessment Framework for Contactless Palmprint Recognition
Contactless palmprint recognition has gained traction due to its convenience and hygienic benefits. However, in real-world scenarios with complex backgrounds and varying hand poses, evaluating image quality to enhance recognition performance remains a significant challenge. To address this, we propose a novel Hand-Geometry Aware Contactless Palmprint Image Quality Assessment (HGAIQA) framework. Unlike existing methods that assess only the palmprint ROI, our framework evaluates the entire image. Firstly, it employs a high-resolution hand segmentation network and keypoint heatmap module to identify hand region and joint keypoints. Secondly, it evaluates the palm’s flatness based on geometric features and assesses additional quality attributes such as brightness and sharpness. Lastly, it determines image quality by analyzing the intra-class and inter-class distributions of fused multi-features. After integrating with subsequent ROI localization and recognition algorithms, experiments show a substantial 21.2% reduction in EER for palmprint recognition on the COEP database by removing the lowest 10% of low-quality images. These results demonstrate the effectiveness of our approach in significantly enhancing palmprint recognition performance.