scholarly journals BUILDING ROOF VECTORIZATION WITH PPGNET

2021 ◽  
Vol XLVI-4/W4-2021 ◽  
pp. 85-90
Author(s):  
S. Hensel ◽  
S. Goebbels ◽  
M. Kada
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Feasibility Study ◽  
Dual Graph ◽  
Post Processing ◽  
Building Roof ◽  
Point Pair ◽  
Edge Detectors ◽  
Initial Results

Abstract. A challenge in data-based 3D building reconstruction is to find the exact edges of roof facet polygons. Although these edges are visible in orthoimages, convolution-based edge detectors also find many other edges due to shadows and textures. In this feasibility study, we apply machine learning to solve this problem. Recently, neural networks have been introduced that not only detect edges in images, but also assemble the edges into a graph. When applied to roof reconstruction, the vertices of the dual graph represent the roof facets. In this study, we apply the Point-Pair Graph Network (PPGNet) to orthoimages of buildings and evaluate the quality of the detected edge graphs. The initial results are promising, and adjusting the training parameters further improved the results. However, in some cases, additional work, such as post-processing, is required to reliably find all vertices.

scholarly journals Prediction Skill of Extended Range 2-m Maximum Air Temperature Probabilistic Forecasts Using Machine Learning Post-Processing Methods

Atmosphere ◽  
2020 ◽  
Vol 11 (8) ◽  
pp. 823
Author(s):  
Ting Peng ◽  
Xiefei Zhi ◽  
Yan Ji ◽  
Luying Ji ◽  
Ye Tian
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Air Temperature ◽  
Post Processing ◽  
Learning Methods ◽  
Processing Methods ◽  
Machine Learning Methods ◽  
Extended Range ◽  
The Neural Networks ◽  
Maximum Air Temperature

The extended range temperature prediction is of great importance for public health, energy and agriculture. The two machine learning methods, namely, the neural networks and natural gradient boosting (NGBoost), are applied to improve the prediction skills of the 2-m maximum air temperature with lead times of 1–35 days over East Asia based on the Environmental Modeling Center, Global Ensemble Forecast System (EMC-GEFS), under the Subseasonal Experiment (SubX) of the National Centers for Environmental Prediction (NCEP). The ensemble model output statistics (EMOS) method is conducted as the benchmark for comparison. The results show that all the post-processing methods can efficiently reduce the prediction biases and uncertainties, especially in the lead week 1–2. The two machine learning methods outperform EMOS by approximately 0.2 in terms of the continuous ranked probability score (CRPS) overall. The neural networks and NGBoost behave as the best models in more than 90% of the study area over the validation period. In our study, CRPS, which is not a common loss function in machine learning, is introduced to make probabilistic forecasting possible for traditional neural networks. Moreover, we extend the NGBoost model to atmospheric sciences of probabilistic temperature forecasting which obtains satisfying performances.

scholarly journals Quantification of the suitable rooftop area for solar panel installation from overhead imagery using Convolutional Neural Networks

2021 ◽  
Vol 2042 (1) ◽  
pp. 012002
Author(s):  
Roberto Castello ◽  
Alina Walch ◽  
Raphaël Attias ◽  
Riccardo Cadei ◽  
Shasha Jiang ◽  
...  
Keyword(s):  
Neural Network ◽  
Machine Learning ◽  
Neural Networks ◽  
Computer Vision ◽  
State Of The Art ◽  
Solar Panel ◽  
Post Processing ◽  
Processing Step ◽  
Recent Method

Abstract The integration of solar technology in the built environment is realized mainly through rooftop-installed panels. In this paper, we leverage state-of-the-art Machine Learning and computer vision techniques applied on overhead images to provide a geo-localization of the available rooftop surfaces for solar panel installation. We further exploit a 3D building database to associate them to the corresponding roof geometries by means of a geospatial post-processing approach. The stand-alone Convolutional Neural Network used to segment suitable rooftop areas reaches an intersection over union of 64% and an accuracy of 93%, while a post-processing step using building database improves the rejection of false positives. The model is applied to a case study area in the canton of Geneva and the results are compared with another recent method used in the literature to derive the realistic available area.

scholarly journals Approach to Define the Reliability of Safety-Related Machine Learning Based Functions in Highly Automated Driving

2021 ◽  
Author(s):  
Pengyu Si ◽  
Ossmane Krini ◽  
Nadine Müller ◽  
Aymen Ouertani
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Pedestrian Detection ◽  
Quality Criteria ◽  
Automated Driving ◽  
Safety Requirements ◽  
Functional Reliability ◽  
Current Standards ◽  
Highly Automated Driving

Current standards cannot cover the safety requirements of machine learning based functions used in highly automated driving. Because of the opacity of neural networks, some self-driving functions cannot be developed following the V-model. These functions require the expansion of the standards. This paper focuses on this gap and defines functional reliability for such functions to help the future standards control the quality of machine learning based functions. As an example, reliability functions for pedestrian detection are built. Since the quality criteria in computer vision do not consider safety, new approaches for expression and evaluation of this reliability are designed.

scholarly journals ALS POINT CLOUD CLASSIFICATION USING POINTNET++ AND KPCONV WITH PRIOR KNOWLEDGE

2021 ◽  
Vol XLVI-4/W4-2021 ◽  
pp. 91-96
Author(s):  
M. Kada ◽  
D. Kuramin
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Prior Knowledge ◽  
Laser Scanning ◽  
Point Clouds ◽  
Machine Learning Algorithms ◽  
Supervised Machine Learning ◽  
Software Applications ◽  
Point Cloud Classification

Abstract. In the practical and professional work of classifying airborne laser scanning (ALS) point clouds, there are nowadays numerous methods and software applications available that are able to separate the points into a few basic categories and do so with a known and consistent quality. Further refinement of the classes then requires either manual or semi-automatic work, or the use of supervised machine learning algorithms. In using supervised machine learning, e.g. Deep Learning neural networks, however, there is a significant chance that they will not maintain the approved quality of an existing classification. In this study, we therefore evaluate the application of two neural networks, PointNet++ and KPConv, and propose to integrate prior knowledge from a pre-existing classification in the form of height above ground and an encoding of the already available labels as additional per-point input features. Our experiments show that such an approach can improve the quality of the 3D classification results by 6% to 10% in mean intersection over union (mIoU) depending on the respective network, but it also cannot completely avoid the aforementioned problems.

Machine Learning to Perform Segmentation and 3D Projection of Abnormal Tissues by Endoscopy Images

2020 ◽  
Vol 17 (5) ◽  
pp. 2296-2303
Author(s):  
V. Adithya Pothan Raj ◽  
P. Mohan Kumar
Keyword(s):  
Machine Learning ◽  
Normal Direction ◽  
Iteration Algorithm ◽  
Post Processing ◽  
Edge Detector ◽  
Performance Outcome ◽  
Direction Estimation ◽  
Segmentation Algorithms ◽  
Anatomical Parameters

Images obtained by endoscopy technique provides the normal direction of the tissue contour. This provides the important anatomical parameters which can be used for segmentation algorithms. Due to the variation of tissue image sizes, the values of intensity for the tissues is typically ununiformed and also have noisiness by nature. So, identifying the direction in normal by a single iteration is unreliable. A multi (factor)-iteration algorithm has been developed for estimating the direction normal to the edge of defective tissue. From experimented results, the estimation reliability is formulated by multiple iterations. The estimation post last iteration corrects the direction normally. We have obtained the balance at all points during the normal direction estimation and it is used by the Edge Detector. The implementation results obtained prove that our proposed algorithm reduces the amount of astonishing boundaries and gapes in the actual outlines. Thus improves the quality of segmentation and 3D projection. The obtained corrected output could also be used in the removal of false edges in post processing. The performance outcome of our proposed algorithm is measured at multiple iterations and results are tabulated.

Application of Artificial Neural Networks for Processing and Interpretation of Data from a Scanning Magnetic Introscope

2021 ◽  
Author(s):  
Victor Evgenevich Kosarev ◽  
Ekaterina Anatolevna Yachmeneva ◽  
Aleksandr Vladimirovich Starovoyto ◽  
Dmitrii Ivanovich Kirgizov ◽  
Rustem Ramilevich Mukhamadiev ◽  
...  
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Artificial Neural Networks ◽  
Machine Learning Algorithms ◽  
Test Machine ◽  
Subjective Factor ◽  
Mathematical Algorithms ◽  
Speed Up ◽  
Artificial Neural

Summary This paper presents the efficiency of using artificial neural networks for solving problems of processing and interpreting geophysical data obtained by scanning magnetic introscopy. Neural networks of various architectures have been implemented to solve the problems of processing primary material, searching for well structure objects,identifying casing defects. The analysis of the capabilities of neural networks in comparison with mathematical algorithms is carried out. To test machine learning algorithms and mathematical algorithms for processing, visualizing and storing the results, a software shell was created in which all tasks are solved using a set of tools. It was found that the use of artificial neural networks can significantly speed up the process of data processing and interpretation, as well as improve the quality of the results in comparison with individual mathematical algorithms. Nevertheless, the use of mathematical algorithms in solving some problems gives consistently better results. In particular, the problematic aspects were identified at the stage of interpretation when identifying defects. This is due to the presence of conventions in the isolation of defects by the operator at the stage of preparing data for training neural networks, which is a subjective factor and requires a deeper study.

scholarly journals The phonetic value of the Proto-Indo-European laryngeals

2021 ◽  
pp. 1-59
Author(s):  
Frederik Hartmann
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Deep Neural Networks ◽  
Learning Algorithms ◽  
Machine Learning Algorithms ◽  
New Method ◽  
Phonetic Features ◽  
Sound Features ◽  
Phonetic Value

Abstract Discussion of the exact phonetic value of the so-called ‘laryngeals’ in Proto-Indo-European has been ongoing ever since their discovery, and no uniform consensus has yet been reached. This paper aims at introducing a new method to determine the quality of the laryngeals that differs substantially from traditional techniques previously applied to this problem, by making use of deep neural networks as part of the larger field of machine learning algorithms. Phonetic environment data serves as the basis for training the networks, enabling the algorithm to determine sound features solely by their immediate phonetic neighbors. It proves possible to assess the phonetic features of the laryngeals computationally and to propose a quantitatively founded interpretation.

scholarly journals An Optimum Tea Fermentation Detection Model Based on Deep Convolutional Neural Networks

Data ◽  
2020 ◽  
Vol 5 (2) ◽  
pp. 44
Author(s):  
Gibson Kimutai ◽  
Alexander Ngenzi ◽  
Rutabayiro Ngoga Said ◽  
Ambrose Kiprop ◽  
Anna Förster
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Machine Learning Techniques ◽  
Support Vector ◽  
K Nearest Neighbor ◽  
Deep Convolutional Neural Networks ◽  
Linear Discriminant ◽  
Learning Techniques ◽  
Classification Tasks

Tea is one of the most popular beverages in the world, and its processing involves a number of steps which includes fermentation. Tea fermentation is the most important step in determining the quality of tea. Currently, optimum fermentation of tea is detected by tasters using any of the following methods: monitoring change in color of tea as fermentation progresses and tasting and smelling the tea as fermentation progresses. These manual methods are not accurate. Consequently, they lead to a compromise in the quality of tea. This study proposes a deep learning model dubbed TeaNet based on Convolution Neural Networks (CNN). The input data to TeaNet are images from the tea Fermentation and Labelme datasets. We compared the performance of TeaNet with other standard machine learning techniques: Random Forest (RF), K-Nearest Neighbor (KNN), Decision Tree (DT), Support Vector Machine (SVM), Linear Discriminant Analysis (LDA), and Naive Bayes (NB). TeaNet was more superior in the classification tasks compared to the other machine learning techniques. However, we will confirm the stability of TeaNet in the classification tasks in our future studies when we deploy it in a tea factory in Kenya. The research also released a tea fermentation dataset that is available for use by the community.

Sign in / Sign up

Export Citation Format

Share Document