Classification and Separation of Diffraction Energy on Pre-Migration Seismic Data using Deep Learning

Diffractions are a useful aspect of the seismic wavefield and are often underutilised. By separating the diffractions from the rest of the wavefield they can be used for various applications such as velocity analysis, structural imaging, and wavefront tomography. However, separating the diffractions is a challenging task due to the comparatively low amplitudes of diffractions as well as the overlap between reflection and diffraction energy. Whilst there are existing analytical methods for separation, these act to remove reflections, leaving a volume which contains diffractions and noise. On top of this, analytical separation techniques can be costly computationally as well as requiring manual parameterisation. To alleviate these issues, a deep neural network has been trained to automatically identify and separate diffractions from reflections and noise on pre-migration data.Here, a Generative Adversarial Network (GAN) has been trained for the automated separation. This is a type of deep neural network architecture which contains two neural networks which compete against one another. One neural network acts as a generator, creating new data which appears visually similar to the real data, while a second neural network acts as a discriminator, trying to identify whether the given data is real or fake. As the generator improves, so too does the discriminator, giving a deeper understanding of the data. To avoid overfitting to a specific dataset as well as to improve the cross-data applicability of the network, data from several different seismic datasets from geologically distinct locations has been used in training. When comparing a network trained on a single dataset compared to one trained on several datasets, it is seen that providing additional data improves the separation on both the original and new datasets.The automatic separation technique is then compared with a conventional, analytical, separation technique; plane-wave destruction (PWD). The computational cost of the GAN separation is vastly superior to that of PWD, performing a separation in minutes on a 3-D dataset in comparison to hours. Although in some complex areas the GAN separation is of a higher quality than the PWD separation, as it does not rely on the dip, there are also areas where the PWD outperforms the GAN separation. The GAN may be enhanced by adding more training data as well as by improving the initial separation used to create the training data, which is based around PWD and thus is imperfect and can introduce bias into the network. A potential for this is training the GAN entirely using synthetic data, which allows for a perfect separation as the points are known, however, it must be of sufficient volume for training and sufficient quality for real data applicability.

Download Full-text

Method of determination of the text direction on the image with the use of convolutional neural network

Informatization and communication ◽

10.34219/2078-8320-2020-11-2-96-99 ◽

2020 ◽

pp. 96-99

Author(s):

P.L. Nikolaev

Keyword(s):

Neural Network ◽

Convolutional Neural Network ◽

Deep Neural Network ◽

Binary Classification ◽

Synthetic Data ◽

Real Data ◽

Method Of Determination ◽

Classification Of Images

This article deals with method of binary classification of images with small text on them Classification is based on the fact that the text can have 2 directions – it can be positioned horizontally and read from left to right or it can be turned 180 degrees so the image must be rotated to read the sign. This type of text can be found on the covers of a variety of books, so in case of recognizing the covers, it is necessary first to determine the direction of the text before we will directly recognize it. The article suggests the development of a deep neural network for determination of the text position in the context of book covers recognizing. The results of training and testing of a convolutional neural network on synthetic data as well as the examples of the network functioning on the real data are presented.

Download Full-text

Robust Approach to Supervised Deep Neural Network Training for Real-Time Object Classification in Cluttered Indoor Environment

Applied Sciences ◽

10.3390/app11157148 ◽

2021 ◽

Vol 11 (15) ◽

pp. 7148

Author(s):

Bedada Endale ◽

Abera Tullu ◽

Hayoung Shi ◽

Beom-Soo Kang

Keyword(s):

Neural Network ◽

Deep Learning ◽

Real Time ◽

Network Architecture ◽

Input Data ◽

Deep Neural Network ◽

Data Augmentation ◽

Object Classification ◽

Training Data ◽

Gradient Descent Algorithm

Unmanned aerial vehicles (UAVs) are being widely utilized for various missions: in both civilian and military sectors. Many of these missions demand UAVs to acquire artificial intelligence about the environments they are navigating in. This perception can be realized by training a computing machine to classify objects in the environment. One of the well known machine training approaches is supervised deep learning, which enables a machine to classify objects. However, supervised deep learning comes with huge sacrifice in terms of time and computational resources. Collecting big input data, pre-training processes, such as labeling training data, and the need for a high performance computer for training are some of the challenges that supervised deep learning poses. To address these setbacks, this study proposes mission specific input data augmentation techniques and the design of light-weight deep neural network architecture that is capable of real-time object classification. Semi-direct visual odometry (SVO) data of augmented images are used to train the network for object classification. Ten classes of 10,000 different images in each class were used as input data where 80% were for training the network and the remaining 20% were used for network validation. For the optimization of the designed deep neural network, a sequential gradient descent algorithm was implemented. This algorithm has the advantage of handling redundancy in the data more efficiently than other algorithms.

Download Full-text

DISIR: DEEP IMAGE SEGMENTATION WITH INTERACTIVE REFINEMENT

ISPRS Annals of Photogrammetry Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-annals-v-2-2020-877-2020 ◽

2020 ◽

Vol V-2-2020 ◽

pp. 877-884

Author(s):

G. Lenczner ◽

B. Le Saux ◽

N. Luminari ◽

A. Chan-Hon-Tong ◽

G. Le Besnerais

Keyword(s):

Neural Network ◽

Network Architecture ◽

Deep Neural Network ◽

Training Data ◽

Aerial Images ◽

Interactive Approach ◽

Rgb Images ◽

Deep Image ◽

Class Segmentation ◽

The Impact

Abstract. This paper presents an interactive approach for multi-class segmentation of aerial images. Precisely, it is based on a deep neural network which exploits both RGB images and annotations. Starting from an initial output based on the image only, our network then interactively refines this segmentation map using a concatenation of the image and user annotations. Importantly, user annotations modify the inputs of the network – not its weights – enabling a fast and smooth process. Through experiments on two public aerial datasets, we show that user annotations are extremely rewarding: each click corrects roughly 5000 pixels. We analyze the impact of different aspects of our framework such as the representation of the annotations, the volume of training data or the network architecture. Code is available at this address.

Download Full-text

Strut Diameter Uncertainty Prediction by Deep Neural Network for Additively Manufactured Lattice Structures

10.1115/detc2021-69985 ◽

2021 ◽

Author(s):

Recep M. Gorguluarslan ◽

Gorkem Can Ates ◽

O. Utku Gungor ◽

Yusuf Yamaner

Keyword(s):

Neural Network ◽

Computer Aided Design ◽

Deep Neural Network ◽

Computational Cost ◽

Training Data ◽

Model Parameters ◽

Effective Diameter ◽

Lattice Structures ◽

Validation Data ◽

Geometric Uncertainties

Abstract Additive manufacturing introduces geometric uncertainties on the fabricated strut members of lattice structures. These uncertainties lead to deviations between the simulation result and the fabricated mechanical performance. Although these uncertainties can be characterized and quantified in the existing literature, the generation of a high number of samples for the quantified uncertainties to use in the computer-aided design of lattice structures for different strut diameters and angles requires high experimental effort and computational cost. The use of deep neural network models to accurately predict the samples of uncertainties is studied in this research to address this issue. For the training data, the geometric uncertainties on the fabricated struts introduced by the material extrusion process are characterized from microscope measurements using random field theory. These uncertainties are propagated to effective diameters of the strut members using a stochastic upscaling technique. The relationship between the deterministic strut model parameters, namely the model diameter and angle, and the effective diameter with propagated uncertainties is established through a deep neural network model. The validation data results show accurate predictions for the effective diameter when model parameters are given as inputs. Thus, the proposed model has the potential to use the fabricated results in the design optimization processes without requiring computationally expensive repetitive simulations.

Download Full-text

Road images augmentation with synthetic traffic signs using neural networks

Computer Optics ◽

10.18287/2412-6179-co-859 ◽

2021 ◽

Vol 5 (45) ◽

pp. 736-748

Author(s):

A.S. Konushin ◽

B.V. Faizov ◽

V.I. Shakhuro

Keyword(s):

Synthetic Data ◽

Real Data ◽

Training Data ◽

Traffic Sign ◽

Generative Adversarial Network ◽

Traffic Signs ◽

Adversarial Network ◽

Sign Recognition ◽

Variational Autoencoder ◽

Sign Detection

Traffic sign recognition is a well-researched problem in computer vision. However, the state of the art methods works only for frequent sign classes, which are well represented in training datasets. We consider the task of rare traffic sign detection and classification. We aim to solve that problem by using synthetic training data. Such training data is obtained by embedding synthetic images of signs in the real photos. We propose three methods for making synthetic signs consistent with a scene in appearance. These methods are based on modern generative adversarial network (GAN) architectures. Our proposed methods allow realistic embedding of rare traffic sign classes that are absent in the training set. We adapt a variational autoencoder for sampling plausible locations of new traffic signs in images. We demonstrate that using a mixture of our synthetic data with real data improves the accuracy of both classifier and detector.

Download Full-text

RESEARCH OF GENERATIVE ADVERSARIAL NETWORKS FOR THE SYNTHESIS OF NEW MEDICAL DATA

Automation and modeling in design and management of ◽

10.30987/2658-6436-2020-2-17-23 ◽

2020 ◽

Vol 2020 (2) ◽

pp. 17-23

Author(s):

Vladislav Laptev ◽

Vyacheslav Danilov ◽

Olga Gerget

Keyword(s):

Neural Network ◽

Network Architecture ◽

Medical Data ◽

Training Data ◽

Generative Adversarial Networks ◽

Training Procedure ◽

Neural Network Architecture ◽

Generative Adversarial Network ◽

Adversarial Network ◽

Adversarial Networks

The paper considers the development of a Generative Adversarial Network (GAN) for the synthesis of new medical data. The developed GAN consists of two models trained simultaneously: a generative model (G - Generator), estimating the distribution of data, and a discriminating model (D - Discriminator), which estimates the probability that the sample is obtained from the training data, and not from generator G. To create G, we used own neural network architecture based on convolutional layers using experimental functions of Tensor Flow Addons. To create discriminator D, we used a Transfer Learning (TL) approach. The training procedure is to maximize the likelihood that discriminator D will make a mistake. Experiments show that the proposed GAN architecture completely copes with the task of synthesizing of new medical data.

Download Full-text

Accurate and Transferable Multitask Prediction of Chemical Properties with an Atoms-in-Molecule Neural Network

10.26434/chemrxiv.7151435.v2 ◽

2018 ◽

Author(s):

Roman Zubatyuk ◽

Justin S. Smith ◽

Jerzy Leszczynski ◽

Olexandr Isayev

Keyword(s):

Neural Network ◽

Molecular System ◽

Computational Cost ◽

Chemical Properties ◽

The State ◽

Molecular Properties ◽

Training Data ◽

Dft Methods ◽

Benchmark Datasets ◽

Quantum Phenomena

Atomic and molecular properties could be evaluated from the fundamental Schrodinger’s equation and therefore represent different modalities of the same quantum phenomena. Here we present AIMNet, a modular and chemically inspired deep neural network potential. We used AIMNet with multitarget training to learn multiple modalities of the state of the atom in a molecular system. The resulting model shows on several benchmark datasets the state-of-the-art accuracy, comparable to the results of orders of magnitude more expensive DFT methods. It can simultaneously predict several atomic and molecular properties without an increase in computational cost. With AIMNet we show a new dimension of transferability: the ability to learn new targets utilizing multimodal information from previous training. The model can learn implicit solvation energy (like SMD) utilizing only a fraction of original training data, and archive MAD error of 1.1 kcal/mol compared to experimental solvation free energies in MNSol database.

Download Full-text