scholarly journals USING SIMULATION DATA FROM GAMING ENVIRONMENTS FOR TRAINING A DEEP LEARNING ALGORITHM ON 3D POINT CLOUDS

2021 ◽  
Vol VIII-4/W2-2021 ◽  
pp. 67-74
Author(s):  
S. Spiegel ◽  
J. Chen
Keyword(s):  
Neural Networks ◽  
Real World ◽  
Resource Constraints ◽  
Learning Algorithm ◽  
Point Clouds ◽  
Training Data ◽  
Lidar Data ◽  
Learning Approaches ◽  
Data Set ◽  
3D Point Clouds

Abstract. Deep neural networks (DNNs) and convolutional neural networks (CNNs) have demonstrated greater robustness and accuracy in classifying two-dimensional images and three-dimensional point clouds compared to more traditional machine learning approaches. However, their main drawback is the need for large quantities of semantically labeled training data sets, which are often out of reach for those with resource constraints. In this study, we evaluated the use of simulated 3D point clouds for training a CNN learning algorithm to segment and classify 3D point clouds of real-world urban environments. The simulation involved collecting light detection and ranging (LiDAR) data using a simulated 16 channel laser scanner within the the CARLA (Car Learning to Act) autonomous vehicle gaming environment. We used this labeled data to train the Kernel Point Convolution (KPConv) and KPConv Segmentation Network for Point Clouds (KP-FCNN), which we tested on real-world LiDAR data from the NPM3D benchmark data set. Our results showed that high accuracy can be achieved using data collected in a simulator.

scholarly journals Performance of Machine Learning Algorithms and Diversity in Data

2018 ◽  
Vol 210 ◽  
pp. 04019 ◽  
Author(s):  
Hyontai SUG
Keyword(s):  
Machine Learning ◽  
Neural Networks ◽  
Real World ◽  
Learning Algorithms ◽  
Machine Learning Algorithms ◽  
Training Data ◽  
Real World Data ◽  
Random Data ◽  
Data Set ◽  
World Data

Recent world events in go games between human and artificial intelligence called AlphaGo showed the big advancement in machine learning technologies. While AlphaGo was trained using real world data, AlphaGo Zero was trained using massive random data, and the fact that AlphaGo Zero won AlphaGo completely revealed that diversity and size in training data is important for better performance for the machine learning algorithms, especially in deep learning algorithms of neural networks. On the other hand, artificial neural networks and decision trees are widely accepted machine learning algorithms because of their robustness in errors and comprehensibility respectively. In this paper in order to prove that diversity and size in data are important factors for better performance of machine learning algorithms empirically, the two representative algorithms are used for experiment. A real world data set called breast tissue was chosen, because the data set consists of real numbers that is very good property for artificial random data generation. The result of the experiment proved the fact that the diversity and size of data are very important factors for better performance.

scholarly journals AN EMPRICAL POINT ERROR MODEL FOR TLS DERIVED POINT CLOUDS

2016 ◽  
Vol XLI-B5 ◽  
pp. 557-563 ◽  
Author(s):  
Mustafa Ozendi ◽  
Devrim Akca ◽  
Hüseyin Topan
Keyword(s):  
Real World ◽  
Incidence Angle ◽  
A Priori ◽  
Laser Scanner ◽  
Point Clouds ◽  
Error Model ◽  
Random Errors ◽  
Data Set ◽  
3D Point Clouds

The random error pattern of point clouds has significant effect on the quality of final 3D model. The magnitude and distribution of random errors should be modelled numerically. This work aims at developing such an anisotropic point error model, specifically for the terrestrial laser scanner (TLS) acquired 3D point clouds. A priori precisions of basic TLS observations, which are the range, horizontal angle and vertical angle, are determined by predefined and practical measurement configurations, performed at real-world test environments. A priori precision of horizontal (𝜎<sub>𝜃</sub>) and vertical (𝜎<sub>𝛼</sub>) angles are constant for each point of a data set, and can directly be determined through the repetitive scanning of the same environment. In our practical tests, precisions of the horizontal and vertical angles were found as 𝜎<sub>𝜃</sub>=±36.6<sup>𝑐𝑐</sup> and 𝜎<sub>𝛼</sub>=±17.8<sup>𝑐𝑐</sup>, respectively. On the other hand, a priori precision of the range observation (𝜎<sub>𝜌</sub>) is assumed to be a function of range, incidence angle of the incoming laser ray, and reflectivity of object surface. Hence, it is a variable, and computed for each point individually by employing an empirically developed formula varying as 𝜎<sub>𝜌</sub>=±2−12 𝑚𝑚 for a FARO Focus X330 laser scanner. This procedure was followed by the computation of error ellipsoids of each point using the law of variance-covariance propagation. The direction and size of the error ellipsoids were computed by the principal components transformation. The usability and feasibility of the model was investigated in real world scenarios. These investigations validated the suitability and practicality of the proposed method.

scholarly journals A sparsity preserving genetic algorithm for extracting diverse functional 3D designs from deep generative neural networks

2020 ◽  
Vol 6 ◽  
Author(s):  
James D. Cunningham ◽  
Dule Shu ◽  
Timothy W. Simpson ◽  
Conrad S. Tucker
Keyword(s):  
Genetic Algorithm ◽  
Neural Networks ◽  
Point Cloud ◽  
Point Clouds ◽  
3D Models ◽  
Training Data ◽  
Alternative Methods ◽  
Data Set ◽  
Design Exploration ◽  
Latent Space

Generative neural networks (GNNs) have successfully used human-created designs to generate novel 3D models that combine concepts from disparate known solutions, which is an important aspect of design exploration. GNNs automatically learn a parameterization (or latent space) of a design space, as opposed to alternative methods that manually define a parameterization. However, GNNs are typically not evaluated using an explicit notion of physical performance, which is a critical capability needed for design. This work bridges this gap by proposing a method to extract a set of functional designs from the latent space of a point cloud generating GNN, without sacrificing the aforementioned aspects of a GNN that are appealing for design exploration. We introduce a sparsity preserving cost function and initialization strategy for a genetic algorithm (GA) to optimize over the latent space of a point cloud generating autoencoder GNN. We examine two test cases, an example of generating ellipsoid point clouds subject to a simple performance criterion and a more complex example of extracting 3D designs with a low coefficient of drag. Our experiments show that the modified GA results in a diverse set of functionally superior designs while maintaining similarity to human-generated designs in the training data set.

scholarly journals AN EMPRICAL POINT ERROR MODEL FOR TLS DERIVED POINT CLOUDS

2016 ◽  
Vol XLI-B5 ◽  
pp. 557-563
Author(s):  
Mustafa Ozendi ◽  
Devrim Akca ◽  
Hüseyin Topan
Keyword(s):  
Real World ◽  
Incidence Angle ◽  
A Priori ◽  
Laser Scanner ◽  
Point Clouds ◽  
Error Model ◽  
Random Errors ◽  
Data Set ◽  
3D Point Clouds

The random error pattern of point clouds has significant effect on the quality of final 3D model. The magnitude and distribution of random errors should be modelled numerically. This work aims at developing such an anisotropic point error model, specifically for the terrestrial laser scanner (TLS) acquired 3D point clouds. A priori precisions of basic TLS observations, which are the range, horizontal angle and vertical angle, are determined by predefined and practical measurement configurations, performed at real-world test environments. A priori precision of horizontal (𝜎<sub>𝜃</sub>) and vertical (𝜎<sub>𝛼</sub>) angles are constant for each point of a data set, and can directly be determined through the repetitive scanning of the same environment. In our practical tests, precisions of the horizontal and vertical angles were found as 𝜎<sub>𝜃</sub>=±36.6<sup>𝑐𝑐</sup> and 𝜎<sub>𝛼</sub>=±17.8<sup>𝑐𝑐</sup>, respectively. On the other hand, a priori precision of the range observation (𝜎<sub>𝜌</sub>) is assumed to be a function of range, incidence angle of the incoming laser ray, and reflectivity of object surface. Hence, it is a variable, and computed for each point individually by employing an empirically developed formula varying as 𝜎<sub>𝜌</sub>=±2−12 𝑚𝑚 for a FARO Focus X330 laser scanner. This procedure was followed by the computation of error ellipsoids of each point using the law of variance-covariance propagation. The direction and size of the error ellipsoids were computed by the principal components transformation. The usability and feasibility of the model was investigated in real world scenarios. These investigations validated the suitability and practicality of the proposed method.

scholarly journals Towards real-time object detection on edge with deep neural networks

2018 ◽  
Author(s):  
◽  
Zhi Zhang
Keyword(s):  
Neural Network ◽  
Neural Networks ◽  
Object Detection ◽  
Real Time ◽  
Real World ◽  
Deep Neural Networks ◽  
Learning Algorithm ◽  
Cloud Services ◽  
Training Data ◽  
Detection Model

Despite being a core topic for more than several decades, object detection is still receiving increasing attentions due to its irreplaceable importance in a wide variety of applications. Abundant object detectors based on deep neural networks have shown significantly revamped accuracies in recent years. However, it's still the day one for these models to be effectively deployed to real world. In this dissertation, we focus on object detection models which tackle real world problems that are unavailable few years ago. We also aim at making object detectors on the go, which means detectors are not longer required to be run on workstations and cloud services which is latency unfriendly. To achieve these goals, we addressed the problem in two phases: application and deployment. We have done thoughtful research on both areas. Our contribution involves inter-frame information fusing, model knowledge distillation, advanced model flow control for progressive inference, and hardware oriented model design and optimization. More specifically, we proposed a novel cross-frame verification scheme for spatial temporal fused object detection model for sequential images and videos in a proposal and reject favor. To compress model from a learning basis and resolve domain specific training data shortage, we improved the learning algorithm to handle insufficient labeled data by searching for optimal guidance paths from pre-trained models. To further reduce model inference cost, we designed a progressive neural network which run in flexible cost enabled by RNN style decision controller during runtime. We recognize the awkward model deployment problem, especially for object detection models that require excessive customized layers. In response, we propose to use end-to-end neural network which use pure neural network components to substitute traditional post-processing operations. We also applied operator decomposition and graph level and on-device optimization towards real-time object detection on low power edge devices. All these works have achieved state-of-the-art performances and converted to successful applications.

scholarly journals ABSent: Cross-Lingual Sentence Representation Mapping with Bidirectional GANs

2020 ◽  
Vol 34 (05) ◽  
pp. 7756-7763 ◽  
Author(s):  
Zuohui Fu ◽  
Yikun Xian ◽  
Shijie Geng ◽  
Yingqiang Ge ◽  
Yuting Wang ◽  
...  
Keyword(s):  
Neural Networks ◽  
Transfer Learning ◽  
Real World ◽  
Source Code ◽  
Training Data ◽  
Learning Approaches ◽  
Word Level ◽  
Parallel Data ◽  
Parallel Text ◽  
Cross Lingual

A number of cross-lingual transfer learning approaches based on neural networks have been proposed for the case when large amounts of parallel text are at our disposal. However, in many real-world settings, the size of parallel annotated training data is restricted. Additionally, prior cross-lingual mapping research has mainly focused on the word level. This raises the question of whether such techniques can also be applied to effortlessly obtain cross-lingually aligned sentence representations. To this end, we propose an Adversarial Bi-directional Sentence Embedding Mapping (ABSent) framework, which learns mappings of cross-lingual sentence representations from limited quantities of parallel data. The experiments show that our method outperforms several technically more powerful approaches, especially under challenging low-resource circumstances. The source code is available from https://github.com/zuohuif/ABSent along with relevant datasets.

scholarly journals Information-Theoretic Generalization Bounds for Meta-Learning and Applications

Entropy ◽  
2021 ◽  
Vol 23 (1) ◽  
pp. 126
Author(s):  
Sharu Theresa Jose ◽  
Osvaldo Simeone
Keyword(s):  
Learning Algorithm ◽  
Broad Class ◽  
Performance Measure ◽  
Training Data ◽  
Learning To Learn ◽  
Data Set ◽  
Information Theoretic ◽  
Meta Learning ◽  
Task Training ◽  
Test Sets

Meta-learning, or “learning to learn”, refers to techniques that infer an inductive bias from data corresponding to multiple related tasks with the goal of improving the sample efficiency for new, previously unobserved, tasks. A key performance measure for meta-learning is the meta-generalization gap, that is, the difference between the average loss measured on the meta-training data and on a new, randomly selected task. This paper presents novel information-theoretic upper bounds on the meta-generalization gap. Two broad classes of meta-learning algorithms are considered that use either separate within-task training and test sets, like model agnostic meta-learning (MAML), or joint within-task training and test sets, like reptile. Extending the existing work for conventional learning, an upper bound on the meta-generalization gap is derived for the former class that depends on the mutual information (MI) between the output of the meta-learning algorithm and its input meta-training data. For the latter, the derived bound includes an additional MI between the output of the per-task learning procedure and corresponding data set to capture within-task uncertainty. Tighter bounds are then developed for the two classes via novel individual task MI (ITMI) bounds. Applications of the derived bounds are finally discussed, including a broad class of noisy iterative algorithms for meta-learning.

scholarly journals Localization of Scattering Objects Using Neural Networks

Sensors ◽  
2020 ◽  
Vol 21 (1) ◽  
pp. 11
Author(s):  
Domonkos Haffner ◽  
Ferenc Izsák
Keyword(s):  
Neural Networks ◽  
Plane Waves ◽  
Measurement Data ◽  
Scattered Wave ◽  
Training Data ◽  
Data Set ◽  
Main Compound ◽  
Incident Plane ◽  
The Neural Networks ◽  
Simulation Package

The localization of multiple scattering objects is performed while using scattered waves. An up-to-date approach: neural networks are used to estimate the corresponding locations. In the scattering phenomenon under investigation, we assume known incident plane waves, fully reflecting balls with known diameters and measurement data of the scattered wave on one fixed segment. The training data are constructed while using the simulation package μ-diff in Matlab. The structure of the neural networks, which are widely used for similar purposes, is further developed. A complex locally connected layer is the main compound of the proposed setup. With this and an appropriate preprocessing of the training data set, the number of parameters can be kept at a relatively low level. As a result, using a relatively large training data set, the unknown locations of the objects can be estimated effectively.

scholarly journals BLAINDER—A Blender AI Add-On for Generation of Semantically Labeled Depth-Sensing Data

Sensors ◽  
2021 ◽  
Vol 21 (6) ◽  
pp. 2144
Author(s):  
Stefan Reitmann ◽  
Lorenzo Neumann ◽  
Bernhard Jung
Keyword(s):  
Point Clouds ◽  
Training Data ◽  
Sensor Data ◽  
Generation Process ◽  
Data Generation ◽  
Depth Sensor ◽  
Depth Sensing ◽  
Depth Sensors ◽  
3D Point Clouds ◽  
Wide Range

Common Machine-Learning (ML) approaches for scene classification require a large amount of training data. However, for classification of depth sensor data, in contrast to image data, relatively few databases are publicly available and manual generation of semantically labeled 3D point clouds is an even more time-consuming task. To simplify the training data generation process for a wide range of domains, we have developed the BLAINDER add-on package for the open-source 3D modeling software Blender, which enables a largely automated generation of semantically annotated point-cloud data in virtual 3D environments. In this paper, we focus on classical depth-sensing techniques Light Detection and Ranging (LiDAR) and Sound Navigation and Ranging (Sonar). Within the BLAINDER add-on, different depth sensors can be loaded from presets, customized sensors can be implemented and different environmental conditions (e.g., influence of rain, dust) can be simulated. The semantically labeled data can be exported to various 2D and 3D formats and are thus optimized for different ML applications and visualizations. In addition, semantically labeled images can be exported using the rendering functionalities of Blender.

scholarly journals Comparative Analysis of Different Mobile LiDAR Mapping Systems for Ditch Line Characterization

2021 ◽  
Vol 13 (13) ◽  
pp. 2485
Author(s):  
Yi-Chun Lin ◽  
Raja Manish ◽  
Darcy Bullock ◽  
Ayman Habib
Keyword(s):  
Flow Direction ◽  
Sediment Accumulation ◽  
Digital Terrain Model ◽  
Vertical Direction ◽  
Point Clouds ◽  
Lidar Data ◽  
Cross Sectional ◽  
Premature Failure ◽  
3D Point Clouds ◽  
Mapping Systems

Maintenance of roadside ditches is important to avoid localized flooding and premature failure of pavements. Scheduling effective preventative maintenance requires a reasonably detailed mapping of the ditch profile to identify areas in need of excavation to remove long-term sediment accumulation. This study utilizes high-resolution, high-quality point clouds collected by mobile LiDAR mapping systems (MLMS) for mapping roadside ditches and performing hydrological analyses. The performance of alternative MLMS units, including an unmanned aerial vehicle, an unmanned ground vehicle, a portable backpack system along with its vehicle-mounted version, a medium-grade wheel-based system, and a high-grade wheel-based system, is evaluated. Point clouds from all the MLMS units are in agreement within the ±3 cm range for solid surfaces and ±7 cm range for vegetated areas along the vertical direction. The portable backpack system that could be carried by a surveyor or mounted on a vehicle is found to be the most cost-effective method for mapping roadside ditches, followed by the medium-grade wheel-based system. Furthermore, a framework for ditch line characterization is proposed and tested using datasets acquired by the medium-grade wheel-based and vehicle-mounted portable systems over a state highway. An existing ground-filtering approach—cloth simulation—is modified to handle variations in point density of mobile LiDAR data. Hydrological analyses, including flow direction and flow accumulation, are applied to extract the drainage network from the digital terrain model (DTM). Cross-sectional/longitudinal profiles of the ditch are automatically extracted from the LiDAR data and visualized in 3D point clouds and 2D images. The slope derived from the LiDAR data turned out to be very close to the highway cross slope design standards of 2% on driving lanes, 4% on shoulders, and a 6-by-1 slope for ditch lines.

Sign in / Sign up

Export Citation Format

Share Document