Yiming Tang
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
Yiming Tang has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Bo Li
Nanjing University of Posts and Telecommunications
1 shared publication - Li-Jie Xu
College of Electronic and Optical Engineering & College of Flexible Electronics, Nanjing University of Posts and Telecommunications, Nanjing, China
1 shared publication Liqing Xu
School of Electronic Science and Engineering, Nanjing University of Posts and Telecommunications, 210023, China
1 shared publication- Ming Zhang
Zhejiang University, Hangzhou, China
1 shared publication Yaming Bo
School of Electronic Science and Engineering, Nanjing University of Posts and Telecommunications, 210023, China
1 shared publication- Yi Yu
Hiroshima University, Japan
1 shared publication
Latest Publications
- Open Access
- 15 March 2023
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 19 (2s) , 1-20
Query-Guided Prototype Learning with Decoder Alignment and Dynamic Fusion in Few-Shot Segmentation
Few-shot segmentation aims to segment objects belonging to a specific class under the guidance of a few annotated examples. Most existing approaches follow the prototype learning paradigm and generate category prototypes by squeezing masked feature maps extracted from images in the support set. These support prototypes may lead to inaccurate predictions when directly compared with features extracted from the query set due to the considerable distribution discrepancy between support and query features. We propose a query-guided prototype learning architecture to address this problem from two aspects: (i) We propose a cross-alignment loss for training the segmentation decoder. This loss function will help the decoder improve its robustness against the distribution discrepancy between support and query features. (ii) We build a dynamic fusion module to strengthen the original support prototype with another prototype extracted from query features. Experiments show that our method achieves promising results compared to previous prototype learning methods on PASCAL-5 i and COCO-20 i datasets.
- 15 March 2023
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 19 (2s) , 1-20
Open AccessQuery-Guided Prototype Learning with Decoder Alignment and Dynamic Fusion in Few-Shot Segmentation
Few-shot segmentation aims to segment objects belonging to a specific class under the guidance of a few annotated examples. Most existing approaches follow the prototype learning paradigm and generate category prototypes by squeezing masked feature maps extracted from images in the support set. These support prototypes may lead to inaccurate predictions when directly compared with features extracted from the query set due to the considerable distribution discrepancy between support and query features. We propose a query-guided prototype learning architecture to address this problem from two aspects: (i) We propose a cross-alignment loss for training the segmentation decoder. This loss function will help the decoder improve its robustness against the distribution discrepancy between support and query features. (ii) We build a dynamic fusion module to strengthen the original support prototype with another prototype extracted from query features. Experiments show that our method achieves promising results compared to previous prototype learning methods on PASCAL-5 i and COCO-20 i datasets.
- 1 August 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3328
Analysis and implementation of 3D frequency selective structures using multi-layer PCBs
Summary form only given. It is well-known that frequency selective surfaces operate as spatial filters for the incoming electromagnetic waves with different operating frequencies, polarizations, and incident angles [1]. Due to these distinctive characteristics, frequency selective surfaces are widely used in many applications of microwave and antenna systems, such as antenna sub-reflectors, radomes, absorbers, etc. [1]. Recently, a new kind of frequency selective surfaces, classified as 3D frequency selective structures (FSSs), has been reported to achieve superior performance [2-4]. In a 3D FSS, multiple resonators are constructed in each 3D-cavity based unit cell, providing a number of transmission zeros/poles at finite frequencies. Subsequently, it is very easy to obtain high performance (such as high selectivity, wide out-of-band rejection, and multi-band response) by controlling locations of these transmission zeros/poles. However, the main disadvantage of 3D FSSs is their complicated topologies, which will definitely increase the implementation difficulties when achieving high performance. In this work, a brief review of 3D FSS designs is first conducted to understand their operating principles and to identify their advantages and limitations. After that, multi-layer PCB technology is introduced to realize 3D FSSs, which can extremely release implementation difficulties of 3D structures. Two examples of multi-layer based 3D FSSs are proposed, as shown in Fig. 1. Operating principles of both examples are explained with the aids of equivalent circuit models and current distributions. Experimental verifications are also provided for both design examples.
- 1 August 2016
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3328
Analysis and implementation of 3D frequency selective structures using multi-layer PCBs
Summary form only given. It is well-known that frequency selective surfaces operate as spatial filters for the incoming electromagnetic waves with different operating frequencies, polarizations, and incident angles [1]. Due to these distinctive characteristics, frequency selective surfaces are widely used in many applications of microwave and antenna systems, such as antenna sub-reflectors, radomes, absorbers, etc. [1]. Recently, a new kind of frequency selective surfaces, classified as 3D frequency selective structures (FSSs), has been reported to achieve superior performance [2-4]. In a 3D FSS, multiple resonators are constructed in each 3D-cavity based unit cell, providing a number of transmission zeros/poles at finite frequencies. Subsequently, it is very easy to obtain high performance (such as high selectivity, wide out-of-band rejection, and multi-band response) by controlling locations of these transmission zeros/poles. However, the main disadvantage of 3D FSSs is their complicated topologies, which will definitely increase the implementation difficulties when achieving high performance. In this work, a brief review of 3D FSS designs is first conducted to understand their operating principles and to identify their advantages and limitations. After that, multi-layer PCB technology is introduced to realize 3D FSSs, which can extremely release implementation difficulties of 3D structures. Two examples of multi-layer based 3D FSSs are proposed, as shown in Fig. 1. Operating principles of both examples are explained with the aids of equivalent circuit models and current distributions. Experimental verifications are also provided for both design examples.