Article
Engineering, Electrical & Electronic
Yuanqiang Cai, Chang Liu, Peirui Cheng, Dawei Du, Libo Zhang, Weiqiang Wang, Qixiang Ye
Summary: The paper introduces a large-scale scene text detection dataset (LS-Text) and a Scale-residual Learning Network (SLN) to address the scale variation problem. By integrating learnable feature concatenation and feature up-sampling operator, SLN effectively eliminates residuals in both Feature Fusion Residuals (FFR) and Scale Transformation Residuals (STR), improving feature representation and performance.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2021)
Article
Computer Science, Information Systems
Zhengzheng Tu, Yan Ma, Zhun Li, Chenglong Li, Jieming Xu, Yongtao Liu
Summary: Salient object detection in complex scenes is challenging, with RGBT SOD emerging as a new research direction. The VT5000 dataset contains 5000 spatially aligned RGBT image pairs and a baseline approach that utilizes multi-level feature extraction and attention mechanism for accurate RGBT SOD.
IEEE TRANSACTIONS ON MULTIMEDIA
(2023)
Article
Engineering, Electrical & Electronic
Ning Xu, An-An Liu, Yongkang Wong, Weizhi Nie, Yuting Su, Mohan Kankanhalli
Summary: This paper proposes a multi-scale context modeling method for scene graph inference, which jointly discovers and integrates object-centric and region-centric context information. Experimental results show that this method can achieve competitive performance on three benchmarks.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2021)
Review
Plant Sciences
Simon Rouet, Romain Barillot, Denis Leclercq, Marie-Helene Bernicot, Didier Combes, Abraham Escobar-Gutierrez, Jean-Louis Durand
Summary: The reproductive development of perennial grasses is influenced by complex interactions between climatic conditions and genetic diversity, which are crucial for grassland management and understanding the potential impacts of climate change. Reproductive development at an individual tiller scale significantly affects plant perenniality, and existing grassland models have limitations in explaining the complexities of reproductive development and genetic x environmental interactions. Introducing underlying processes involved in reproductive development into models would enhance predictions of grassland behavior under future growth conditions.
FRONTIERS IN PLANT SCIENCE
(2021)
Article
Geochemistry & Geophysics
Ariyan Zarei, Emmanuel Gonzalez, Nirav Merchant, Duke Pauli, Eric Lyons, Kobus Barnard
Summary: This paper addresses the challenge of fast image stitching for large image collections while effectively dealing with drift and minimal overlap. The authors propose a method that focuses on scientific applications and prioritizes ground-truth accuracy. They present approaches for both affine and homography transformations and demonstrate the superiority of their method through evaluations on various datasets. The authors also provide valuable ground-truth datasets for further research in this field.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)
Article
Computer Science, Interdisciplinary Applications
Jiacheng Wang, Fei Chen, Yuxi Ma, Liansheng Wang, Zhaodong Fei, Jianwei Shuai, Xiangdong Tang, Qichao Zhou, Jing Qin
Summary: This paper proposes a novel cross-scale boundary-aware transformer network, XBound-Former, to simultaneously address the variation and boundary problems in skin lesion segmentation. The network captures boundary knowledge through three specially designed learners and performs boundary-aware attention at multiple scales. Experimental results demonstrate that the model consistently outperforms other models on skin lesion datasets and polyp lesion datasets, especially on boundary-wise metrics.
IEEE TRANSACTIONS ON MEDICAL IMAGING
(2023)
Article
Computer Science, Artificial Intelligence
Xianfa Xu, Zhe Chen, Fuliang Yin
Summary: This study presents a monocular depth estimation method with multi-scale spatial attention guidance and semantic enhancement, which can focus more on small objects and improve the sharpness of depth prediction edges. Experimental results on public benchmark datasets demonstrate the effectiveness and superior performance of the proposed method.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2021)
Article
Computer Science, Information Systems
Daniel Padilla Carrasco, Hatem A. Rashwan, Miguel Angel Garcia, Domenec Puig
Summary: In this paper, a modified and lightweight deep object detection model based on the YOLO-v5 architecture is proposed, which can detect objects of various sizes. It employs a multi-scale mechanism to learn deep discriminative feature representations at different scales and automatically determine the most suitable scales for object detection in a scene (i.e., vehicles in this case). Experimental results demonstrate significant improvement in precision compared to the original YOLO-v5 architecture.
Article
Robotics
Lac Van Duong, Van Anh Ho
Summary: The sense of touch is crucial for individuals to interact with their environment, and even more so for robots. TacLINK is a large-scale tactile sensing system that can perceive tactile information through the deformation of its soft skin.
IEEE TRANSACTIONS ON ROBOTICS
(2021)
Article
Horticulture
Yu-Chun Chu, Jer-Chia Chang
Summary: This study used the BBCH scale to investigate the phenological stages of red-fleshed pitaya and recorded the development of spine and flesh color, providing valuable information for scientific research and cultivation practices of red-fleshed pitaya.
SCIENTIA HORTICULTURAE
(2022)
Article
Agronomy
Caner Ferhatoglu, Bradley A. Miller
Summary: This study compared the effectiveness of six types of feature selection methods from four different categories in digital soil mapping. The results showed that wrapper and embedded methods usually led to optimal models, while filter-based methods often resulted in overfit models. Decision-tree models were usually part of the optimal combination of feature selection and machine learning.
Article
Computer Science, Artificial Intelligence
Kaiyang Zhou, Yongxin Yang, Andrea Cavallaro, Tao Xiang
Summary: In this paper, we propose an effective person re-identification (re-ID) model that can distinguish similar-looking people and can be deployed across datasets without any adaptation. The model consists of an omni-scale network (OSNet) for feature learning and instance normalisation (IN) layers for improving generalisation. Experimental results show that the proposed model outperforms existing re-ID models in terms of performance, both in the same-dataset and cross-dataset settings.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Engineering, Electrical & Electronic
Chalavadi Vishnu, Rajeshreddy Datla, Debaditya Roy, Sobhan Babu, C. Krishna Mohan
Summary: The proposed approach efficiently models spatio-temporal features using fall motion vector, achieving better performance in fall detection in various scenarios compared to existing methods.
IEEE SENSORS JOURNAL
(2021)
Article
Environmental Sciences
Marvin Lauenburg, Matthias Karl, Volker Matthias, Markus Quante, Martin Otto Paul Ramacher
Summary: This study examines the dispersion of ultrafine particles (UFP) in Hamburg city center, with a focus on the impact of passenger ferryboats. Using a chemical transport model and observations, the study models and compares particle number concentrations. Results show that ferryboat traffic is a significant source of emissions near the shore, with slight differences between modeled and measured concentrations. UFP concentrations exhibit strong spatial and temporal variations, mainly driven by air temperature, wind speed, and wind direction.
Article
Computer Science, Information Systems
Caihua Liu, Yifan Chen, Xinyu He, Tao Xu
Summary: This research proposes a multi-view crowd counting method, SASNet, which addresses the issue of scale inconsistency across views and the complex background. The experimental results demonstrate that SASNet outperforms existing methods in crowd counting.
Review
Ecology
Nagai Shin, Taku. M. M. Saitoh, Yayoi Takeuchi, Tomoaki Miura, Masahiro Aiba, Hiroko Kurokawa, Yusuke Onoda, Kazuhito Ichii, Kenlo Nishida Nasahara, Rikie Suzuki, Tohru Nakashizuka, Hiroyuki Muraoka
Summary: This article discusses the use of remote sensing observations to evaluate the spatial and temporal variability of land use and cover types and plant phenology in East Asia. Recent advancements in satellite technologies have allowed for high-resolution monitoring of these variables. The authors encourage interdisciplinary collaboration to enhance monitoring capabilities.
ECOLOGICAL RESEARCH
(2023)
Article
Geography, Physical
Yunze Zang, Yuean Qiu, Xuehong Chen, Jin Chen, Wei Yang, Yifei Liu, Longkang Peng, Miaogen Shen, Xin Cao
Summary: RSG-OC is an automated rapeseed mapping approach that combines rule-based sample generation and a one-class classifier (PUL-RF). It generates cloud-free samples during the predicted flowering period using empirical index-based sample selection rules, utilizes all available features based on the rapeseed phenological calendar for classification, and improves generalization to pixels without cloud-free observations using a specific sample augmentation. The method achieved a high accuracy of 94.90% in mapping rapeseed in China.
GISCIENCE & REMOTE SENSING
(2023)
Article
Biodiversity Conservation
Duncan A. O'Brien, Gideon Gal, Stephen J. Thackeray, Shin-ichiro S. Matsuzaki, Christopher F. Clements
Summary: Managing ecosystems requires reliable tools to infer system stability and dynamics. Functional diversity (FD) is an important measure of biodiversity that is conceptually linked to ecological processes. However, it is unclear whether changes in FD occur before or after changes in system state. In this study, we investigated the lagged relationship between planktonic FD and abundance-based metrics in five lake communities. The results show that FD and lake state display synchrony, but the strength of the relationship varies between lakes. Changes in FD may not be identifiable before changes in easily collected abundance metrics. Empirical dynamic modelling is a powerful tool for understanding time lagged relationships in complex ecosystems.
GLOBAL CHANGE BIOLOGY
(2023)
Article
Environmental Sciences
Tomoki Morozumi, Tomomichi Kato, Hideki Kobayashi, Yuma Sakai, Naohisa Nakashima, Kanokrat Buareal, Kenlo Nishida Nasahara, Tomoko Kawaguchi Akitsu, Shohei Murayama, Hibiki M. Noda, Hiroyuki Muraoka
Summary: This study aimed to investigate the seasonal and diurnal variations of ground-based SIF and its representability as total SIF emissions. The results showed that SIF above the overstory and the sum of SIF from the three layers increased sharply during the leaf-onset of the overstory and correlated with each season's GPP. In summary, a multi-layer approach can be used to understand the relationship between SIF and photosynthesis within the canopy.
REMOTE SENSING OF ENVIRONMENT
(2023)
Editorial Material
Environmental Studies
Ioanna Daphne Giannoulatou, Stephanie R. Januchowski-Hartley, Asha Sahni, Sayali K. Pawar, James C. White, Julia Lockheart, Nina Baranduin, Doryn Herbst, Benjamin Whittaker, Stephen J. Thackeray, Robert Shooter, Judy Darley, Merryn Thomas
Summary: This article contributes to creative geography by examining river spaces and a multimodal practice conducted by a group of artists/poets/scientists. The collaboration resulted in two collage series that incorporated visual, textual, and audio-visual media. The article shares and discusses the lessons learned from the group's online creative practice during the global pandemic, including the possibilities of online collaboration, the internet's role in facilitating distant and continuous co-creation, and the potential of collage and poetry in reimagining relationships with rivers.
CULTURAL GEOGRAPHIES
(2023)
Article
Ecology
Christian Korner, Patrick Mohl, Erika Hiltbrunner
Summary: The concept of growing season in terrestrial ecosystems, which determines plant biomass production, lacks a well-defined definition. This study shows different aspects of growing season, including the actual growth period of plants, the period defined by phenological markers, the period of vegetation achieving net primary production, and the potential growth period based on meteorological criteria. The duration of this "window of opportunity" is a strong predictor for global net primary production, especially in forests. These different definitions have important implications for understanding and modeling plant growth and biomass production, challenging the common view that phenology is a proxy for productivity variation.
Article
Environmental Sciences
Yuhao Pan, Dailiang Peng, Jing M. Chen, Ranga B. Myneni, Xiaoyang Zhang, Alfredo R. Huete, Yongshuo H. Fu, Shijun Zheng, Kai Yan, Le Yu, Peng Zhu, Miaogen Shen, Weimin Ju, Wenquan Zhu, Qiaoyun Xie, Wenjiang Huang, Zhengchao Chen, Jingfeng Huang, Chaoyang Wu
Summary: Global warming has caused earlier spring green-up dates, impacting global carbon and hydrologic cycles. Land cover change (LCC) can also affect these dates. By analyzing satellite data from 1992 to 2020, we found that climate variables had a larger impact overall, but LCC controlled the variability in green-up dates in 6% of the Northern Hemisphere. Changes in land cover types led to earlier or later green-up dates in specific regions. Ignoring the impact of LCC overestimated the climate change attribution of earlier green-up dates by three days.
ENVIRONMENTAL RESEARCH LETTERS
(2023)
Article
Ecology
Christian Korner, Gunter Hoch
Summary: This perspective article addresses the recent interest in treeline studies as a result of attempts to identify climate warming effects on mountain and arctic vegetation. The article emphasizes the importance of clear-cut definitions, consistent terminology, and a theoretical framework for hypothesis testing. By applying the ecological niche concept, the potential and realized niche edges can be used to define the climatic limit and deviations of tree growth at treeline. Additionally, it is explained why other abiotic factors, such as microclimate and moisture, do not diminish the classical isotherm concept.
JOURNAL OF BIOGEOGRAPHY
(2023)
Article
Biology
Akira S. Mori, Kureha F. Suzuki, Masakazu Hori, Taku Kadoya, Kotaro Okano, Aya Uraguchi, Hiroyuki Muraoka, Tamotsu Sato, Hideaki Shibata, Yukari Suzuki-Ohno, Keisuke Koba, Mariko Toda, Shin-ichi Nakano, Michio Kondoh, Kaoru Kitajima, Masahiro Nakamura
Summary: As interest in natural capital grows and society increasingly recognizes the value of biodiversity, we must discuss how ecosystem observations to detect changes in biodiversity can be sustained through collaboration across regions and sectors. However, there are many barriers to establishing and sustaining large-scale, fine-resolution ecosystem observations. First, comprehensive monitoring data on both biodiversity and possible anthropogenic factors are lacking. Second, some in situ ecosystem observations cannot be systematically established and maintained across locations. Third, equitable solutions across sectors and countries are needed to build a global network.
PHILOSOPHICAL TRANSACTIONS OF THE ROYAL SOCIETY B-BIOLOGICAL SCIENCES
(2023)
Article
Forestry
Cong Gao, Chunming Shi, Yuxin Lou, Ran An, Cheng Sun, Guocan Wu, Yuandong Zhang, Miaogen Shen, Deliang Chen
Summary: To reconstruct the past Arctic temperature variations, the authors screened 1116 tree-ring width and density records and applied four detrending functions. They selected 338-396 records that showed significant correlations with instrumental temperature data and combined them into a proxy record. The reconstruction explained 45-57% of the instrumental temperature variance since 1950, and the estimated Arctic warming amplitudes were consistent with temperature anomaly datasets.
Article
Meteorology & Atmospheric Sciences
Nan Jiang, Miaogen Shen, Jin Chen, Wei Yang, Xiaolin Zhu, Xufeng Wang, Josep Penuelas
Summary: Previous studies have shown a significant advance in vegetation green-up (VGD) onset date in the Northern Hemisphere during the 1980s and 1990s. However, later studies based on advanced very high-resolution radiometer (AVHRR) data suggested a hiatus in this trend during the warming period from the late 1990s to early 2010s. There is uncertainty in this finding due to quality issues associated with AVHRR data. Our study, using high-quality Moderate Resolution Imaging Spectroradiometer data, shows that VGD advanced significantly despite the warming hiatus, suggesting caution in inferring climate warming based on spring phenology advances.
NPJ CLIMATE AND ATMOSPHERIC SCIENCE
(2023)
Review
Ecology
Nagai Shin, Chifuyu Katsumata, Tomoaki Miura, Narumasa Tsutsumida, Tomoaki Ichie, Ayumi Kotani, Michiko Nakagawa, Kho Lip Khoon, Hideki Kobayashi, Tomo'omi Kumagai, Shunsuke Tei, Runi Anak Sylvester Pungga, Taizo Yamada, Akihiro Kameda, Masayuki Yanagisawa, Kenlo Nishida Nasahara, Hiroyuki Muraoka, Kazuhito Ichii, Yuji Tokumoto
Summary: Recent advances in satellite-borne optical sensors have greatly improved the monitoring of tropical ecosystems in Asia, which have been significantly impacted by human activities and climate change. Using the Multispectral Instrument (MSI) on Sentinel-2A/2B satellites, we can monitor phenology among tree species with a 10-meter spatial resolution and 5-day observational intervals. The Advanced Himawari Imager (AHI) on the Himawari-8 geostationary satellite, with a 1,000-meter spatial resolution and 10-minute observational intervals, can accurately detect the timing and patterns of phenology among tree species using vegetation indices, although the sensor's spatial resolution needs improvement. Geolocation data from text, pictures, and historical field notes published on the Internet can be used to validate land cover and land use.
FRONTIERS IN FORESTS AND GLOBAL CHANGE
(2023)
Article
Biodiversity Conservation
Licong Liu, Jin Chen, Miaogen Shen, Xuehong Chen, Ruyin Cao, Xin Cao, Xihong Cui, Wei Yang, Xiaolin Zhu, Le Li, Yanhong Tang
Summary: We propose a novel method for remotely sensing alpine grasslines and determining their positions, which is of great importance for investigating the response of alpine grasslands to climate change.
GLOBAL CHANGE BIOLOGY
(2023)
Article
Environmental Sciences
Zhiguang Chen, Miaogen Shen, Nan Jiang, Jin Chen, Yanhong Tang, Song Gu
Summary: Daytime warming can delay the end of the vegetation growing season on the Tibetan Plateau, despite the inhibitory effect of low temperatures on alpine vegetation activity. Researchers should take into account the interactive effects of temperature and precipitation on the timing of the growing season when modeling autumn phenology in this region.
SCIENCE OF THE TOTAL ENVIRONMENT
(2023)
Article
Environmental Sciences
Chienwei Tao, Tong Guo, Miaogen Shen, Yanhong Tang
Summary: This study used remote sensing data and the LandTrendr algorithm to analyze the spatio-temporal dynamics of disturbances in planted and natural forests in Northern China. The results show that both types of forests experienced severe disturbances in specific years, with over one third of the forest area being highly disturbed. The duration of disturbances in planted and natural forests was mostly 1 to 3 years. After drought events, the NDVI anomaly of planted forests showed a slow upward variation, while natural forests did not.
