Image Anonymization using Deep Convolutional Generative Adversarial Network

Abstract Advancement in deep learning requires significantly huge amount of data for training purpose, where protection of individual data plays a key role in data privacy and publication. Recent developments in deep learning demonstarte a huge challenge for traditionally used approch for image anonymization, such as model inversion attack, where adversary repeatedly query the model, inorder to reconstrut the original image from the anonymized image. In order to apply more protection on image anonymization, an approach is presented here to convert the input (raw) image into a new synthetic image by applying optimized noise to the latent space representation (LSR) of the original image. The synthetic image is anonymized by adding well designed noise calculated over the gradient during the learning process, where the resultant image is both realistic and immune to model inversion attack. More presicely, we extend the approach proposed by T. Kim and J. Yang, 2019 by using Deep Convolutional Generative Adversarial Network (DCGAN) in order to make the approach more efficient. Our aim is to improve the efficiency of the model by changing the loss function to achieve optimal privacy in less time and computation. Finally, the proposed approach is demonstrated using a benchmark dataset. The experimental study presents that the proposed method can efficiently convert the input image into another synthetic image which is of high quality as well as immune to model inversion attack.

Download Full-text

Synthetic image data augmentation for fibre layup inspection processes: Techniques to enhance the data set

Journal of Intelligent Manufacturing ◽

10.1007/s10845-021-01738-7 ◽

2021 ◽

Author(s):

Sebastian Meister ◽

Nantwin Möller ◽

Jan Stüve ◽

Roger M. Groves

Keyword(s):

Deep Learning ◽

Test Method ◽

Process Efficiency ◽

Synthetic Image ◽

Generative Adversarial Network ◽

Data Set ◽

Manufacturing Defects ◽

Fiber Placement ◽

Adversarial Network ◽

Geometrical Transformation

AbstractIn the aerospace industry, the Automated Fiber Placement process is an established method for producing composite parts. Nowadays the required visual inspection, subsequent to this process, typically takes up to 50% of the total manufacturing time and the inspection quality strongly depends on the inspector. A Deep Learning based classification of manufacturing defects is a possibility to improve the process efficiency and accuracy. However, these techniques require several hundreds or thousands of training data samples. Acquiring this huge amount of data is difficult and time consuming in a real world manufacturing process. Thus, an approach for augmenting a smaller number of defect images for the training of a neural network classifier is presented. Five traditional methods and eight deep learning approaches are theoretically assessed according to the literature. The selected conditional Deep Convolutional Generative Adversarial Network and Geometrical Transformation techniques are investigated in detail, with regard to the diversity and realism of the synthetic images. Between 22 and 166 laser line scan sensor images per defect class from six common fiber placement inspection cases are utilised for tests. The GAN-Train GAN-Test method was applied for the validation. The studies demonstrated that a conditional Deep Convolutional Generative Adversarial Network combined with a previous Geometrical Transformation is well suited to generate a large realistic data set from less than 50 actual input images. The presented network architecture and the associated training weights can serve as a basis for applying the demonstrated approach to other fibre layup inspection images.

Download Full-text

Unpaired Image Denoising via Wasserstein GAN in Low-Dose CT Image with Multi-Perceptual Loss and Fidelity Loss

Symmetry ◽

10.3390/sym13010126 ◽

2021 ◽

Vol 13 (1) ◽

pp. 126

Author(s):

Zhixian Yin ◽

Kewen Xia ◽

Ziping He ◽

Jiangnan Zhang ◽

Sijie Wang ◽

...

Keyword(s):

Computed Tomography ◽

Deep Learning ◽

Image Denoising ◽

Low Dose ◽

Radiation Risk ◽

Semantic Features ◽

Original Image ◽

Generative Adversarial Network ◽

Adversarial Network ◽

The Difference

The use of low-dose computed tomography (LDCT) in medical practice can effectively reduce the radiation risk of patients, but it may increase noise and artefacts, which can compromise diagnostic information. The methods based on deep learning can effectively improve image quality, but most of them use a training set of aligned image pairs, which are difficult to obtain in practice. In order to solve this problem, on the basis of the Wasserstein generative adversarial network (GAN) framework, we propose a generative adversarial network combining multi-perceptual loss and fidelity loss. Multi-perceptual loss uses the high-level semantic features of the image to achieve the purpose of noise suppression by minimizing the difference between the LDCT image and the normal-dose computed tomography (NDCT) image in the feature space. In addition, L2 loss is used to calculate the loss between the generated image and the original image to constrain the difference between the denoised image and the original image, so as to ensure that the image generated by the network using the unpaired images is not distorted. Experiments show that the proposed method performs comparably to the current deep learning methods which utilize paired image for image denoising.

Download Full-text

Network-Simulated Generation of Human Faces with Expressions and Orientations for Deep Learning Classification

International Journal of Innovative Technology and Exploring Engineering - Special Issue ◽

10.35940/ijitee.b7491.129219 ◽

2019 ◽

Vol 9 (2) ◽

pp. 2178-2185

Keyword(s):

Deep Learning ◽

Synthetic Image ◽

Human Face ◽

Medical Field ◽

Generative Adversarial Network ◽

Human Face Recognition ◽

Adversarial Network ◽

Prediction Time ◽

Rapid Prediction ◽

Human Faces

Human face recognition is a complex task, and it is important as it can be applied to assist people worldwide, such as those in the medical field or security. For example, human faces can be used for detecting pain or emotion. Nevertheless, a drawback of deep learning methods that need a lot of data to process is key. In this study, a deep-learning-based technique, which is used to classify, that generates a synthetic image of the facial expression and orientation by utilizing the Wasserstein generative adversarial network (WGAN) is presented. The WGAN can improve the performance of the deep learning method. The proposed system certainly generates images with a small number of datasets compared to the large datasets. This research aims to solve the problem of deep learning by increasing the accuracy of the system. The generated output coincides with the real image dataset. The application using ResNet-50 and RetinaNet as a pre-model for the prediction and detection of the human faces revealed a rapid prediction time and accuracy during the assessment test.

Download Full-text

An Imbalanced Image Classification Method for the Cell Cycle Phase

Information ◽

10.3390/info12060249 ◽

2021 ◽

Vol 12 (6) ◽

pp. 249

Author(s):

Xin Jin ◽

Yuanwen Zou ◽

Zhongbing Huang

Keyword(s):

Cell Cycle ◽

Deep Learning ◽

Image Classification ◽

Classification Accuracy ◽

Data Augmentation ◽

Cycle Phase ◽

Generative Adversarial Network ◽

Adversarial Network ◽

Cellular Life

The cell cycle is an important process in cellular life. In recent years, some image processing methods have been developed to determine the cell cycle stages of individual cells. However, in most of these methods, cells have to be segmented, and their features need to be extracted. During feature extraction, some important information may be lost, resulting in lower classification accuracy. Thus, we used a deep learning method to retain all cell features. In order to solve the problems surrounding insufficient numbers of original images and the imbalanced distribution of original images, we used the Wasserstein generative adversarial network-gradient penalty (WGAN-GP) for data augmentation. At the same time, a residual network (ResNet) was used for image classification. ResNet is one of the most used deep learning classification networks. The classification accuracy of cell cycle images was achieved more effectively with our method, reaching 83.88%. Compared with an accuracy of 79.40% in previous experiments, our accuracy increased by 4.48%. Another dataset was used to verify the effect of our model and, compared with the accuracy from previous results, our accuracy increased by 12.52%. The results showed that our new cell cycle image classification system based on WGAN-GP and ResNet is useful for the classification of imbalanced images. Moreover, our method could potentially solve the low classification accuracy in biomedical images caused by insufficient numbers of original images and the imbalanced distribution of original images.

Download Full-text

AN AI-BASED APPROACH TO ENHANCED FRACTURE RESOLUTION IN IMAGE LOGS

10.30632/spwla-2021-0081 ◽

2021 ◽

Author(s):

James Howard ◽

◽

Joe Tracey ◽

Mike Shen ◽

Shawn Zhang ◽

...

Keyword(s):

Neural Network ◽

Deep Learning ◽

Nearest Neighbor ◽

Rock Fracture ◽

Short Interval ◽

Acoustic Properties ◽

Generative Adversarial Network ◽

Adversarial Network ◽

Deep Learning Neural Network ◽

Borehole Image

Borehole image logs are used to identify the presence and orientation of fractures, both natural and induced, found in reservoir intervals. The contrast in electrical or acoustic properties of the rock matrix and fluid-filled fractures is sufficiently large enough that sub-resolution features can be detected by these image logging tools. The resolution of these image logs is based on the design and operation of the tools, and generally is in the millimeter per pixel range. Hence the quantitative measurement of actual width remains problematic. An artificial intelligence (AI) -based workflow combines the statistical information obtained from a Machine-Learning (ML) segmentation process with a multiple-layer neural network that defines a Deep Learning process that enhances fractures in a borehole image. These new images allow for a more robust analysis of fracture widths, especially those that are sub-resolution. The images from a BHTV log were first segmented into rock and fluid-filled fractures using a ML-segmentation tool that applied multiple image processing filters that captured information to describe patterns in fracture-rock distribution based on nearest-neighbor behavior. The robust ML analysis was trained by users to identify these two components over a short interval in the well, and then the regression model-based coefficients applied to the remaining log. Based on the training, each pixel was assigned a probability value between 1.0 (being a fracture) and 0.0 (pure rock), with most of the pixels assigned one of these two values. Intermediate probabilities represented pixels on the edge of rock-fracture interface or the presence of one or more sub-resolution fractures within the rock. The probability matrix produced a map or image of the distribution of probabilities that determined whether a given pixel in the image was a fracture or partially filled with a fracture. The Deep Learning neural network was based on a Conditional Generative Adversarial Network (cGAN) approach where the probability map was first encoded and combined with a noise vector that acted as a seed for diverse feature generation. This combination was used to generate new images that represented the BHTV response. The second layer of the neural network, the adversarial or discriminator portion, determined whether the generated images were representative of the actual BHTV by comparing the generated images with actual images from the log and producing an output probability of whether it was real or fake. This probability was then used to train the generator and discriminator models that were then applied to the entire log. Several scenarios were run with different probability maps. The enhanced BHTV images brought out fractures observed in the core photos that were less obvious in the original BTHV log through enhanced continuity and improved resolution on fracture widths.

Download Full-text

Generative Adversarial Networks-Based Semi-Supervised Automatic Modulation Recognition for Cognitive Radio Networks

Sensors ◽

10.3390/s18113913 ◽

2018 ◽

Vol 18 (11) ◽

pp. 3913 ◽

Cited By ~ 6

Author(s):

Mingxuan Li ◽

Ou Li ◽

Guangyi Liu ◽

Ce Zhang

Keyword(s):

Deep Learning ◽

Cognitive Radio ◽

Supervised Learning ◽

Rapid Development ◽

Generative Adversarial Networks ◽

Modulation Recognition ◽

Learning Methods ◽

Generative Adversarial Network ◽

Adversarial Network ◽

Automatic Modulation Recognition

With the recently explosive growth of deep learning, automatic modulation recognition has undergone rapid development. Most of the newly proposed methods are dependent on large numbers of labeled samples. We are committed to using fewer labeled samples to perform automatic modulation recognition in the cognitive radio domain. Here, a semi-supervised learning method based on adversarial training is proposed which is called signal classifier generative adversarial network. Most of the prior methods based on this technology involve computer vision applications. However, we improve the existing network structure of a generative adversarial network by adding the encoder network and a signal spatial transform module, allowing our framework to address radio signal processing tasks more efficiently. These two technical improvements effectively avoid nonconvergence and mode collapse problems caused by the complexity of the radio signals. The results of simulations show that compared with well-known deep learning methods, our method improves the classification accuracy on a synthetic radio frequency dataset by 0.1% to 12%. In addition, we verify the advantages of our method in a semi-supervised scenario and obtain a significant increase in accuracy compared with traditional semi-supervised learning methods.

Download Full-text

INFRASTRUCTURE DEGRADATION AND POST-DISASTER DAMAGE DETECTION USING ANOMALY DETECTING GENERATIVE ADVERSARIAL NETWORKS

ISPRS Annals of Photogrammetry Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-annals-v-2-2020-573-2020 ◽

2020 ◽

Vol V-2-2020 ◽

pp. 573-582 ◽

Cited By ~ 1

Author(s):

S. M. Tilon ◽

F. Nex ◽

D. Duarte ◽

N. Kerle ◽

G. Vosselman

Keyword(s):

Deep Learning ◽

Damage Detection ◽

Training Data ◽

Generative Adversarial Networks ◽

Generative Adversarial Network ◽

Essential Information ◽

Urban Scenes ◽

Adversarial Network ◽

Collapsed Buildings ◽

Post Disaster

Abstract. Degradation and damage detection provides essential information to maintenance workers in routine monitoring and to first responders in post-disaster scenarios. Despite advance in Earth Observation (EO), image analysis and deep learning techniques, the quality and quantity of training data for deep learning is still limited. As a result, no robust method has been found yet that can transfer and generalize well over a variety of geographic locations and typologies of damages. Since damages can be seen as anomalies, occurring sparingly over time and space, we propose to use an anomaly detecting Generative Adversarial Network (GAN) to detect damages. The main advantages of using GANs are that only healthy unannotated images are needed, and that a variety of damages, including the never before seen damage, can be detected. In this study we aimed to investigate 1) the ability of anomaly detecting GANs to detect degradation (potholes and cracks) in asphalt road infrastructures using Mobile Mapper imagery and building damage (collapsed buildings, rubble piles) using post-disaster aerial imagery, and 2) the sensitivity of this method against various types of pre-processing. Our results show that we can detect damages in urban scenes at satisfying levels but not on asphalt roads. Future work will investigate how to further classify the found damages and how to improve damage detection for asphalt roads.

Download Full-text

Improving CBCT Quality to CT Level using Deep‐Learning with Generative Adversarial Network

Medical Physics ◽

10.1002/mp.14624 ◽

2020 ◽

Author(s):

Yang Zhang ◽

Ning Yue ◽

Min‐Ying Su ◽

Bo Liu ◽

Yi Ding ◽

...

Keyword(s):

Deep Learning ◽

Generative Adversarial Network ◽

Adversarial Network

Download Full-text

A generative adversarial network for artifact removal in photoacoustic computed tomography with a linear-array transducer

Experimental Biology and Medicine ◽

10.1177/1535370220914285 ◽

2020 ◽

Vol 245 (7) ◽

pp. 597-605 ◽

Cited By ~ 6

Author(s):

Tri Vu ◽

Mucong Li ◽

Hannah Humayun ◽

Yuan Zhou ◽

Junjie Yao

Keyword(s):

Deep Learning ◽

Linear Array ◽

Artifact Removal ◽

Generative Adversarial Network ◽

Transducer Array ◽

Limited Bandwidth ◽

Adversarial Network ◽

Array Transducer ◽

Imaging Speed

With balanced spatial resolution, penetration depth, and imaging speed, photoacoustic computed tomography (PACT) is promising for clinical translation such as in breast cancer screening, functional brain imaging, and surgical guidance. Typically using a linear ultrasound (US) transducer array, PACT has great flexibility for hand-held applications. However, the linear US transducer array has a limited detection angle range and frequency bandwidth, resulting in limited-view and limited-bandwidth artifacts in the reconstructed PACT images. These artifacts significantly reduce the imaging quality. To address these issues, existing solutions often have to pay the price of system complexity, cost, and/or imaging speed. Here, we propose a deep-learning-based method that explores the Wasserstein generative adversarial network with gradient penalty (WGAN-GP) to reduce the limited-view and limited-bandwidth artifacts in PACT. Compared with existing reconstruction and convolutional neural network approach, our model has shown improvement in imaging quality and resolution. Our results on simulation, phantom, and in vivo data have collectively demonstrated the feasibility of applying WGAN-GP to improve PACT’s image quality without any modification to the current imaging set-up. Impact statement This study has the following main impacts. It offers a promising solution for removing limited-view and limited-bandwidth artifact in PACT using a linear-array transducer and conventional image reconstruction, which have long hindered its clinical translation. Our solution shows unprecedented artifact removal ability for in vivo image, which may enable important applications such as imaging tumor angiogenesis and hypoxia. The study reports, for the first time, the use of an advanced deep-learning model based on stabilized generative adversarial network. Our results have demonstrated its superiority over other state-of-the-art deep-learning methods.

Download Full-text

Measuring Traffic Volumes Using an Autoencoder with No Need to Tag Images with Labels

Electronics ◽

10.3390/electronics9050702 ◽

2020 ◽

Vol 9 (5) ◽

pp. 702

Author(s):

Seungbin Roh ◽

Johyun Shin ◽

Keemin Sohn

Keyword(s):

Input Image ◽

Video Frame ◽

Generative Adversarial Network ◽

Detection Algorithms ◽

Adversarial Network ◽

Simpler Algorithm ◽

Proposed Model ◽

Traffic Volumes ◽

Step Algorithm ◽

Almost All

Almost all vision technologies that are used to measure traffic volume use a two-step procedure that involves tracking and detecting. Object detection algorithms such as YOLO and Fast-RCNN have been successfully applied to detecting vehicles. The tracking of vehicles requires an additional algorithm that can trace the vehicles that appear in a previous video frame to their appearance in a subsequent frame. This two-step algorithm prevails in the field but requires substantial computation resources for training, testing, and evaluation. The present study devised a simpler algorithm based on an autoencoder that requires no labeled data for training. An autoencoder was trained on the pixel intensities of a virtual line placed on images in an unsupervised manner. The last hidden node of the former encoding portion of the autoencoder generates a scalar signal that can be used to judge whether a vehicle is passing. A cycle-consistent generative adversarial network (CycleGAN) was used to transform an original input photo of complex vehicle images and backgrounds into a simple illustration input image that enhances the performance of the autoencoder in judging the presence of a vehicle. The proposed model is much lighter and faster than a YOLO-based model, and accuracy of the proposed model is equivalent to, or better than, a YOLO-based model. In measuring traffic volumes, the proposed approach turned out to be robust in terms of both accuracy and efficiency.

Download Full-text