scholarly journals Building Extraction at Scale Using Convolutional Neural Network: Mapping of the United States

2018 ◽  
Vol 11 (8) ◽  
pp. 2600-2614 ◽  
Author(s):  
Hsiuhan Lexie Yang ◽  
Jiangye Yuan ◽  
Dalton Lunga ◽  
Melanie Laverdiere ◽  
Amy Rose ◽  
...  
Keyword(s):  
Neural Network ◽  
United States ◽  
Convolutional Neural Network ◽  
The United States ◽  
Building Extraction ◽  
Network Mapping

scholarly journals Predicting road quality using high resolution satellite imagery: A transfer learning approach

PLoS ONE ◽  
2021 ◽  
Vol 16 (7) ◽  
pp. e0253370
Author(s):  
Ethan Brewer ◽  
Jason Lin ◽  
Peter Kemper ◽  
John Hennin ◽  
Dan Runfola
Keyword(s):  
Neural Network ◽  
United States ◽  
Convolutional Neural Network ◽  
Transfer Learning ◽  
Satellite Imagery ◽  
The United States ◽  
Accurate Information ◽  
Learning Approach ◽  
Road Infrastructure ◽  
Health And Economic Development

Recognizing the importance of road infrastructure to promote human health and economic development, actors around the globe are regularly investing in both new roads and road improvements. However, in many contexts there is a sparsity—or complete lack—of accurate information regarding existing road infrastructure, challenging the effective identification of where investments should be made. Previous literature has focused on overcoming this gap through the use of satellite imagery to detect and map roads. In this piece, we extend this literature by leveraging satellite imagery to estimate road quality and concomitant information about travel speed. We adopt a transfer learning approach in which a convolutional neural network architecture is first trained on data collected in the United States (where data is readily available), and then “fine-tuned” on an independent, smaller dataset collected from Nigeria. We test and compare eight different convolutional neural network architectures using a dataset of 53,686 images of 2,400 kilometers of roads in the United States, in which each road segment is measured as “low”, “middle”, or “high” quality using an open, cellphone-based measuring platform. Using satellite imagery to estimate these classes, we achieve an accuracy of 80.0%, with 99.4% of predictions falling within the actual or an adjacent class. The highest performing base model was applied to a preliminary case study in Nigeria, using a dataset of 1,000 images of paved and unpaved roads. By tailoring our US-model on the basis of this Nigeria-specific data, we were able to achieve an accuracy of 94.0% in predicting the quality of Nigerian roads. A continuous case estimate also showed the ability, on average, to predict road quality to within 0.32 on a 0 to 3 scale (with higher values indicating higher levels of quality).

scholarly journals A Convolutional Neural Network for Spatial Downscaling of Satellite-Based Solar-Induced Chlorophyll Fluorescence (SIFnet)

2021 ◽  
Author(s):  
Johannes Gensheimer ◽  
Alexander Jay Turner ◽  
Philipp Köhler ◽  
Christian Frankenberg ◽  
Jia Chen
Keyword(s):  
Neural Network ◽  
United States ◽  
Chlorophyll Fluorescence ◽  
Convolutional Neural Network ◽  
Spatial Resolution ◽  
Vegetation Index ◽  
Spatial Scales ◽  
The United States ◽  
Cycle Space ◽  
Auxiliary Data

Abstract. Gross primary productivity (GPP) is the sum of leaf photosynthesis and represents a crucial component of the global carbon cycle. Space-borne estimates of GPP typically rely on observable quantities that co-vary with GPP such as vegetation indices using reflectance measurements (e.g., NDVI, NIRv, and kNDVI). Recent work has also utilized measurements of solar-induced chlorophyll fluorescence (SIF) as a proxy for GPP. However, these SIF measurements are typically coarse resolution while many processes influencing GPP occur at fine spatial scales. Here, we develop a Convolutional Neural Network (CNN), named SIFnet, that increases the resolution of SIF from the TROPOspheric Monitoring Instrument (TROPOMI) on board of the satellite Sentinel-5P by a factor of 10 to a spatial resolution of 500 m. SIFnet utilizes coarse SIF observations together with high resolution auxiliary data. The auxiliary data used here may carry information related to GPP and SIF. We use training data from non-US regions between April 2018 until March 2021 and evaluate our CNN over the conterminous United States (CONUS). We show that SIFnet is able to increase the resolution of TROPOMI SIF by a factor of 10 with a r2 and RMSE metrics of 0.92 and 0.17 mW m−2 sr−1 nm−1, respectively. We further compare SIFnet against a recently developed downscaling approach and evaluate both methods against independent SIF measurements from Orbiting Carbon Observatory 2 and 3 (OCO-2/3). SIFnet performs systematically better than the downscaling approach (r = 0.78 for SIFnet, r = 0.72 for downscaling), indicating that it is picking up on key features related to SIF and GPP. Examination of the feature importance in the neural network indicates a few key parameters and the spatial regions these parameters matter. Namely, the CNN finds low resolution SIF data to be the most significant parameter with the NIRv vegetation index as the second most important parameter. NIRv consistently outperforms the recently proposed kNDVI vegetation index. Advantages and limitations of SIFnet are investigated and presented through a series of case studies across the United States. SIFnet represents a robust method to infer continuous, high spatial resolution SIF data.

scholarly journals Klasifikasi Citra Pigmen Kanker Kulit Menggunakan Convolutional Neural Network

2021 ◽  
Vol 5 (2) ◽  
pp. 379-385
Author(s):  
Luqman Hakim ◽  
Zamah Sari ◽  
Handhajani Handhajani
Keyword(s):  
Neural Network ◽  
United States ◽  
Neural Networks ◽  
Skin Cancer ◽  
Convolutional Neural Network ◽  
The United States ◽  
Skin Imaging ◽  
Test And Evaluation ◽  
Skin Cancers ◽  
Fully Connected

Skin cancer is a very common form of cancer that can be found in the United States with annual treatment costs exceeding $ 8 billion. New innovations in the classification and detection of skin cancer using artificial neural networks continue to develop to help the medical and medical world in analyzing images accurately and accurately. Researchers propose to classify skin cancer pigments by focusing on two classes, namely non-melanocytic malignant and benign, where the skin cancer category which is classified into the non-melanocytic class is Actinic keratoses, Basal cell carcinoma. While skin cancers that are classified into Benign are Benign keratosis like lesions, dermatofibrama, vascular lessions. The method used in this study is Convolutional Neural Network (CNN) with a model architecture using 8 Convolutional 2D layers which have filters (16, 16, 32, 32, 64, 64, 128, 128). The first input layers are (20,20). and the following layers (5,5 and 3,3), the types of pooling used in this study are MaxPooling and AveragePooling. The Fully Connected Layer used is (256, 128) and uses a Dropout (0.2). The dataset is obtained from the International Skin Imaging Collaboration (ISIC) 2018 with a total of 10015 images. Based on the results of the test and evaluation reports, an accuracy of 75% is obtained. with the highest precision and recall values ​​found in the Benign class, namely 0.80 and 0.82 respectively and the f1_score value of 0.81.  

scholarly journals The United States’ Clothing Imports from Asian Countries along the Belt and Road: An Extended Gravity Trade Model with Application of Artificial Neural Network

2020 ◽  
Vol 12 (18) ◽  
pp. 7433
Author(s):  
Danny Chi Kuen Ho ◽  
Eve Man Hin Chan ◽  
Tsz Leung Yip ◽  
Chi-Wing Tsang
Keyword(s):  
Neural Network ◽  
United States ◽  
Artificial Neural Network ◽  
Developing Countries ◽  
Trade Flow ◽  
The United States ◽  
Extended Gravity ◽  
Belt And Road ◽  
Artificial Neural ◽  
The Belt And Road

In 2013, China announced the Belt and Road Initiative (BRI), which aims to promote the connectivity of Asia, Europe, and Africa and deepen mutually beneficial economic cooperation among member countries. Past studies have reported a positive impact of the BRI on trade between China and its partner countries along the Belt and Road (B&R). However, less is known about its effect on the sectoral trade between the B&R countries and countries that show little support of the BRI. To address that gap, this study examines the changing patterns of clothing imports by the United States (US) from China and 14 B&R countries in Asia. An extended gravity model with a policy variable BRI is built to explain bilateral clothing trade flow. A panel regression model and artificial neural network (ANN) are developed based on the data collected from 1998 to 2018 and applied to predict the trade pattern of 2019. The results show a positive effect of the BRI on the clothing exports of some Asian developing countries along the B&R to the US and demonstrate the superior predictive power of the ANN. More research is needed to examine the balance between economic growth and the social and environmental sustainability of developing countries and to apply more advanced machine learning algorithms to examine global trade flow under the BRI.

Transport Energy Demand Modeling of the United States Using Artificial Neural Networks and Multiple Linear Regressions

2014 ◽  
Author(s):  
Arash Kialashaki ◽  
John Reisel
Keyword(s):  
Neural Network ◽  
United States ◽  
Energy Demand ◽  
Regression Models ◽  
The United States ◽  
Transportation Energy ◽  
Independent Variables ◽  
Transportation Sector ◽  
The Future ◽  
Artificial Neural

In 2009, the transportation sector was the second largest consumer of primary energy in the United States, following the electric power sector and followed by the industrial, residential, and commercial sectors. The pattern of energy use varies by sector. For example, petroleum provides 96% of the energy used for transportation but its share is much less in other sectors. While the United States consumes vast quantities of energy, it has also pledged to cut its greenhouse gas emissions by 2050. In order to assist in planning for future energy needs, the purpose of this study is to develop a model for transport energy demand that incorporates past trends. This paper describes the development of two types of transportation energy models which are able to predict the United States’ future transportation energy-demand. One model uses an artificial neural network technique (a feed-forward multilayer perceptron neural network coupled with back-propagation technique), and the other model uses a multiple linear regression technique. Various independent variables (including GDP, population, oil price, and number of vehicles) are tested. The future transport energy demand can then be forecast based on the application of the growth rate of effective parameters on the models. The future trends of independent variables have been predicted based on the historical data from 1980 using a regression method. Using the forecast of independent variables, the energy demand has been forecasted for period of 2010 to 2030. In terms of the forecasts generated, the models show two different trends despite their performances being at the same level during the model-test period. Although, the results from the regression models show a uniform increase with different slopes corresponding to different models for energy demand in the near future, the results from ANN express no significant change in demand in same time frame. Increased sensitivity of the ANN models to the recent fluctuations caused by the economic recession may be the reason for the differences with the regression models which predict based on the total long-term trends. Although a small increase in the energy demand in the transportation sector of the United States has been predicted by the models, additional factors need to be considered regarding future energy policy. For example, the United States may choose to reduce energy consumption in order to reduce CO2 emissions and meet its national and international commitments, or large increases in fuel efficiency may reduce petroleum demand.

scholarly journals Explaining COVID-19 Outbreaks with Reactive SEIRD Models

2021 ◽  
Author(s):  
Kunal Menda ◽  
Lucas Laird ◽  
Mykel J. Kochenderfer ◽  
Rajmonda S. Caceres
Keyword(s):  
Neural Network ◽  
United States ◽  
Infection Rate ◽  
Expectation Maximization ◽  
The United States ◽  
County Level ◽  
Training Set ◽  
Single Model ◽  
Partial Observability ◽  
Level Data

AbstractCOVID-19 epidemics have varied dramatically in nature across the United States, where some counties have clear peaks in infections, and others have had a multitude of unpredictable and non-distinct peaks. In this work, we seek to explain the diversity in epidemic progressions by considering an extension to the compartmental SEIRD model. The model we propose uses a neural network to predict the infection rate as a function of time and of the prevalence of the disease. We provide a methodology for fitting this model to available county-level data describing aggregate cases and deaths. Our method uses Expectation-Maximization in order to overcome the challenge of partial observability—that the system’s state is only partially reflected in available data. We fit a single model to data from multiple counties in the United States exhibiting different behavior. By simulating the model, we show that it is capable of exhibiting both single peak and multi-peak behavior, reproducing behavior observed in counties both in and out of the training set. We also numerically compare the error of simulations from our model with a standard SEIRD model, showing that the proposed extensions are necessary to be able to explain the spread of COVID-19.

Sign in / Sign up

Export Citation Format

Share Document