Article
Instruments & Instrumentation
Lichun Mei, Caiyun Wang, Huaiye Wang, Yuanfu Zhao, Jun Zhang, Xiaoxia Zhao
Summary: This paper proposes a scheme called Hybrid Matching by Pixel Distribution Mapping (HMPDM) that combines traditional methods and neural network methods to achieve fast template matching. The scheme overcomes the challenge of nonlinear intensity differences between multi-modal images by extracting and mapping pixel distribution information. The experimental results demonstrate the real-time and accurate performance of the scheme.
INFRARED PHYSICS & TECHNOLOGY
(2022)
Article
Engineering, Electrical & Electronic
Lichun Mei, Yuanfu Zhao, Huaiye Wang, Caiyun Wang, Jun Zhang, Xiaoxia Zhao
Summary: This article introduces an efficient and accurate method for template matching in multisource images by combining traditional algorithms and neural networks to reduce the use of system resources. The experimental results show that this method achieves good matching performance with limited hardware resources.
IET SIGNAL PROCESSING
(2023)
Article
Engineering, Electrical & Electronic
Pingcheng Dong, Zhuoyu Chen, Zhuoao Li, Yuzhe Fu, Lei Chen, Fengwei An
Summary: This paper proposes a hardware-oriented SGM algorithm with pixel-level pipeline and region-optimized cost aggregation for high-speed processing and low hardware-resource usage. The algorithm is demonstrated on low-cost XILINX Spartan-7 and advanced Stratix-V FPGA devices for VGA depth estimation, achieving high processing speeds and energy efficiency.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS I-REGULAR PAPERS
(2022)
Article
Computer Science, Information Systems
Shuhua Ma, Peikai Guo, Hairong You, Ping He, Guanglin Li, Heng Li
Summary: This paper introduces a matching optimization algorithm called PSC-RANSAC, which enhances image matching accuracy by using density peaks clustering to select mismatches and effectively eliminating residual mismatches compared to other algorithms.
INFORMATION SCIENCES
(2021)
Article
Chemistry, Analytical
Yunhao Ma, Xiwei Fang, Xinyu Guan, Ke Li, Lei Chen, Fengwei An
Summary: This paper proposes an improved disparity strategy for better accuracy in binocular stereoscopic matching, and designs a hardware architecture based on optimization algorithms. Experimental results demonstrate the superior performance of this method in long-range applications.
Article
Automation & Control Systems
Jingchun Zhou, Tongyu Yang, Weishen Chu, Weishi Zhang
Summary: This paper proposes a restoration method based on backscatter pixel prior and color cast removal for underwater images. Experimental results show that the proposed method improves the contrast and removes color deviation caused by light absorption.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
(2022)
Article
Environmental Sciences
Buge Liang, Zhenghong Peng, Degui Yang, Xing Wang, Jin Li
Summary: This study proposes a novel translational compensation algorithm based on template matching to address the difficulties in extracting micro-motion features caused by the high-speed movement of ballistic targets in midcourse. The algorithm achieves high precision and low time complexity with lower requirements for time-frequency resolution, and it is also applicable to spectral aliasing.
Article
Computer Science, Information Systems
Shengyu Gao, Hongyu Wang, Xin Lou
Summary: This paper introduces a new end-to-end pipeline for stereo matching from raw Bayer pattern images to disparity maps with customized ISP. By introducing a subsampling-based demosaicing-downsampling operation, unnecessary ISP steps can be skipped and the computational complexity of the pipeline is significantly reduced. The proposed pipeline is capable of generating comparable or even better stereo matching results than traditional pipelines while reducing input image size.
Article
Engineering, Electrical & Electronic
Yeongmin Lee, Hyeji Kim
Summary: This paper introduces a new cluster-wise cost aggregation algorithm and its optimized architecture for implementing a high-throughput semiglobal matching stereo depth estimation algorithm. The proposed approach processes each group of pixels, reducing the timing constraint on recursion through pipelining and parallel computation, and achieves higher efficiency with a memory reduction scheme and a system-on-chip tiled processing scheme.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2022)
Article
Environmental Sciences
Tian Hui, Yuelei Xu, Qing Zhou, Chaofeng Yuan, Jarhinbek Rasol
Summary: This paper proposes a method for cross-view remote sensing image matching to address the challenges introduced by heterogeneous image sources, different scales, and different viewpoints. The method includes the introduction of a spatial attention map to deal with domain gap, a multi-scale matching method for different flight altitudes, and a pixel-wise consensus method for adaptability to viewpoint changes. The proposed model is trained based on weakly supervised learning and evaluation demonstrates its robustness and effectiveness. The method is applicable to three types of template matching with different viewpoints.
Article
Chemistry, Analytical
Youngmo Han
Summary: Template matching is a simple image detection algorithm that can easily detect different types of objects by changing the template without tedious training procedures. However, traditional template matching is not very reliable for images that differ from the template. By utilizing additional information such as depths from the vision sensor system, the reliability of template matching can be improved.
Article
Physics, Multidisciplinary
Ching-Hsun Tseng, Shin-Jye Lee, Jianan Feng, Shengzhong Mao, Yu-Ping Wu, Jia-Yu Shang, Xiao-Jun Zeng
Summary: This work proposes an efficient and robust backbone, UPANets, which utilizes channel and spatial direction attentions to expand the receptive fields in shallow convolutional layers. Experimental results show that UPANets achieve better performance with fewer resources on CIFAR-{10, 100} than existing state-of-the-art methods.
Article
Ecology
Moritz D. Luerig
Summary: Digital images are widely used by biologists to capture and analyze organismal phenotypes. phenopype is a high-throughput phenotyping pipeline for Python that aims to extract high-dimensional phenotypic data from digital images efficiently. It provides functions for image preprocessing, segmentation, data extraction, visualization, and data export, facilitating scientific image analysis for biologists with little coding experience. This software helps in increasing the speed and reproducibility of data collection, making it a valuable tool for researchers in ecological, evolutionary, and developmental biology.
METHODS IN ECOLOGY AND EVOLUTION
(2022)
Article
Computer Science, Information Systems
Sinem Guemues, Fatih Kamisli
Summary: This paper proposes a learned compression system for lossless image compression, achieving state-of-the-art performance with only 59K parameters, much less than other recent learned systems. The system utilizes a neural network to process each pixel's causal neighborhood and obtain probability distribution parameters for compression. Parallel decoding algorithms are implemented to reduce decoding time. The system is compared to traditional and learned systems in terms of compression performance, encoding-decoding times, and computational complexity.
MULTIMEDIA TOOLS AND APPLICATIONS
(2023)
Article
Green & Sustainable Science & Technology
Zainab R. Alhalalmeh, Yasser M. Fouda, Muhammad A. Rushdi, Moawwad El-Mikkawy, Hao-Chiang Koong Lin
Summary: This research addresses the need to improve template-matching performance in e-learning and automated assessments in Egypt's evolving educational landscape, especially during the COVID-19 pandemic. The study proposes enhanced schemes by integrating different feature descriptors and evaluates their effectiveness against a baseline approach. The results suggest that the proposed schemes have the potential to improve personalization, engagement, and learning outcomes in e-learning.