A5101774331- Robot Manipulation and Learning
- Video Surveillance and Tracking Methods
- Indoor and Outdoor Localization Technologies
- Sugarcane Cultivation and Processing
- Robotics and Sensor-Based Localization
- Irrigation Practices and Water Management
- Optical Wireless Communication Technologies
- Land Use and Ecosystem Services
- Natural Products and Biological Research
- Adversarial Robustness in Machine Learning
- Geochemistry and Geologic Mapping
- Advanced Image Processing Techniques
- Robotic Path Planning Algorithms
- Geochemistry and Geochronology of Asian Mineral Deposits
- Modular Robots and Swarm Intelligence
- Reinforcement Learning in Robotics
- Geological and Geochemical Analysis
- UAV Applications and Optimization
- Advanced Vision and Imaging
- Advanced Sensor and Energy Harvesting Materials
- Rice Cultivation and Yield Improvement
- Advanced Data Compression Techniques
- Autophagy in Disease and Therapy
- Visual Attention and Saliency Detection
- Biofuel production and bioconversion
Queensland University of Technology
2015-2024
University of Guelph
2022-2024
Huazhong University of Science and Technology
2024
Alibaba Group (China)
2017-2022
University of Sheffield
2021-2022
Nanjing University of Finance and Economics
2019-2022
State Key Laboratory of Continental Dynamics
2020
University of Florida
2019-2020
China-Japan Friendship Hospital
2020
Northwest University
2020
Siamese network based trackers formulate tracking as convolutional feature cross-correlation between target template and searching region. However, still have accuracy gap compared with state-of-the-art algorithms they cannot take advantage of from deep networks, such ResNet-50 or deeper. In this work we prove the core reason comes lack strict translation invariance. By comprehensive theoretical analysis experimental validations, break restriction through a simple yet effective spatial aware...
Abstract Laser‐induced graphene (LIG) has attracted extensive attention owing to its facile preparation of and direct engraving patterns for devices. Various applications are demonstrated such as sensors, supercapacitors, electrocatalysis, batteries, antimicrobial, oil water separation, solar cells, heaters. In recent years, doping been employed a significant strategy modulate the properties LIG thereby improve performance Due patternable manufacture, controllable morphologies, synergistic...
Robotic challenges like the Amazon Picking Challenge (APC) or DARPA Challenges are an established and important way to drive scientific progress. They make research comparable on a well-defined benchmark with equal test conditions for all participants. However, such challenge events occur only occasionally, limited small number of contestants, very difficult replicate after main event. We present new physical robotic picking: ACRV Benchmark. Designed be reproducible, it consists set 42...
We propose to use visible-light beacons for low-cost indoor localization. Modulated light-emitting diode (LED) lights are adapted localization as well illumination. The proposed solution consists of two components: light-signal decomposition and Bayesian
Various approaches have been proposed to learn visuo-motor policies for real-world robotic applications. One solution is first learning in simulation then transferring the real world. In transfer, most existing need images with labels. However, labeling process often expensive or even impractical many this article, we introduce an adversarial discriminative sim-to-real transfer approach reduce amount of labeled data required. The effectiveness demonstrated modular networks a table-top...
While deep learning has had significant successes in computer vision thanks to the abundance of visual data, collecting sufficiently large real-world datasets for robot can be costly. To increase practicality these techniques on real robots, we propose a modular reinforcement method capable transferring models trained simulation robotic task. We introduce bottleneck between perception and control, enabling networks independently, but then merged fine-tuned an end-to-end manner further...
For mobile robots and position-based services, such as healthcare service, precise localization is the most fundamental capability while low-cost solutions are with increasing need potentially have a wide market. A solution based on novel Visible Light Communication (VLC) system for indoor environments proposed in this paper. number of modulated LED lights used beacons to aid additional illumination. Gaussian Process(GP) model intensity distributions light sources. Bayesian framework...
Siamese network based trackers formulate tracking as convolutional feature cross-correlation between target template and searching region. However, still have accuracy gap compared with state-of-the-art algorithms they cannot take advantage of from deep networks, such ResNet-50 or deeper. In this work we prove the core reason comes lack strict translation invariance. By comprehensive theoretical analysis experimental validations, break restriction through a simple yet effective spatial aware...
Abstract Louisiana is the second largest producer of sugarcane ( Saccharum officinarum L.) in United States. New cultivars better adapted for have been development since inception a three‐way cooperative breeding program 1920s. Piecewise regression used to determine growth both and sugar yields how new mechanization had positive impact yields. The analysis uses several breakpoints periods between 1911 2018 consider varietal mechanization. In most cases, changes significantly increased yield...
Compression standards have been used to reduce the cost of image storage and transmission for decades. In recent years, learned compression methods proposed achieved compelling performance traditional standards. However, in these methods, a set different networks are various rates, resulting high model training. Although some variable-rate approaches by using single network, most them brought degradation when applying fine rate control. To enable control without sacrificing performance, we...
Ethanolic fermentation is frequently performed under conditions of low nitrogen. In Saccharomyces cerevisiae, nitrogen limitation induces macroautophagy, including the selective removal mitochondria, also called mitophagy. Previous research showed that blocking mitophagy by deletion mitophagy-specific gene ATG32 increased performance during brewing Ginjo sake. this study, we tested if a similar strategy could enhance alcoholic in context fuel ethanol production from sugarcane Brazilian...
The continuous growth of the economy and population have promoted increasing consumption natural resources, raised concerns regarding upper limits terrestrial ecosystems with biomass accessible for humanity. Here, human appropriation net primary production (HANPP) was employed to assess influence activities on (NPP), a detailed method introduced simulate magnitude trends HANPP in Yangtze River Delta. results showed that total Delta increased from 102.3 Tg C yr−1 142.2 yr−1, during 2005–2015,...
Land use change is one of the crucial factors affecting carbon emissions. The continuously increasing CO 2 and global warming have raised concerns about emission reduction in process urbanization. In this research, Markov multi-objective optimization models were conducted to predict demands for land Nanjing 2030 2060 under natural growth (NG) minimum (MCE) scenarios coordinate needs economic development, ecological protection food security as well target emissions future. spatial...
Net primary productivity (NPP) is the key component of terrestrial carbon cycle, and NPP trends under increasing CO2 climate change in past future are great significance study global budget. Here, LPJ-DGVM was employed to simulate magnitude pattern China’s using long-term series data understand response concentration change. The results showed that total ecosystem increased from 2.8 3.6 Pg C yr−1 over period 1961–2016, with an annual average 3.1 yr−1. a gradient decrease southeast northwest....
Florida is the largest producer of sugarcane ( Saccharum spp.) in United States. Since 1911, University continues to develop new cultivars that are better adapted environment. We use piecewise regression determine growth both and sugar yields, how have positively affected yields. breakpoints for periods between 1950 2018: Scenario 1 (1950–2018), 2 (1950–1980 1981–2018), 3 (1950–1980, 1981–2000, 2001–2018), 4 (1950–1968, 1969–2000, 2001–2018). In most scenarios, significantly increased yield,...
Face clustering has attracted rising research interest recently to take advantage of massive amounts face images on the web. State-of-the-art performance been achieved by Graph Convolutional Networks (GCN) due their powerful representation capacity. However, existing GCN-based methods build graphs mainly according kNN relations in feature space, which may lead a lot noise edges connecting two faces different classes. The features will be polluted when messages pass along these edges, thus...
Various approaches have been proposed to learn visuo-motor policies for real-world robotic applications. One solution is first learning in simulation then transferring the real world. In transfer, most existing need images with labels. However, labelling process often expensive or even impractical many this paper, we propose an adversarial discriminative sim-to-real transfer approach reduce cost of data. The effectiveness demonstrated modular networks a table-top object reaching task where 7...
Indoor localization is a fundamental capability for service robots and indoor applications on mobile devices. To realize that, the cost performance are of great concern. In this paper, we introduce lightweight signal encoding decomposition method low-cost low-power Visible Light Communication (VLC)-based system. Firstly, Gold-sequence-based tiny-length code selection introduced light encoding. Then correlation-based asynchronous blind light-signal developed lights mixed with modulated...
The framework of dominant learned video compression methods is usually composed motion prediction modules as well vector and residual image modules, suffering from its complex structure error propagation problem. Approaches have been proposed to reduce the complexity by replacing with implicit flow networks. Error aware training strategy also alleviate incremental reconstruction errors previously decoded frames. Although these brought some improvement, little attention has paid itself....
Abstract This analysis investigates the impact of cultivar development and cultural practices on Louisiana rice ( Oryza sativa L.) yield growth using Chow test (Chow, 1960) in combination with regression analysis. The examines 1895–2019 period, uses four scenarios which one, two, three, subperiods show varying rates Louisiana. There has been a significant increase yields over time. However, for some is greater than others. Our results time correspond events such as establishment Louisiana's...
For mobile robots and position-based services, localization is the most fundamental capability while path-planning an important application based on that. A novel solution a low-cost Visible Light Communication (VLC) system for indoor environments proposed in this paper. number of modulated LED lights are used as beacons to aid additional illumination. Gaussian Process (GP) model intensity distributions light sources. Path-planning hereby feasible by using GP variance field, rather than...