scholarly journals A Generative Adversarial Inpainting Network to Enhance Prediction of Periodontal Clinical Attachment Level

2021 ◽  
Author(s):  
Vasant Kearney ◽  
Alfa-Ibrahim M. Yansane ◽  
Ryan G. Brandon ◽  
Ram Vaderhobli ◽  
Guo-Hao Lin ◽  
...  
Keyword(s):  
Prediction Accuracy ◽  
Absolute Error ◽  
Generative Adversarial Networks ◽  
Mean Bias Error ◽  
Bias Error ◽  
X Rays ◽  
Clinical Attachment Level ◽  
Adversarial Networks ◽  
Attachment Level ◽  
Pairwise Test

Abstract Deep learning algorithms has recently been used to determine clinical attachment levels (CAL) which aid in the diagnosis of periodontal disease. However, the limited field-of-view of dental bitewing x-rays poses a challenge for convolutional neural networks (CNN) because out-of-view anatomy cannot be directly considered. This study presents an inpainting algorithm using generative adversarial networks (GANs) coupled with partial convolutions to predict out-of-view anatomy to enhance CAL prediction accuracy. 80,326 images were used for training, 12,901 images were used for validation and 10,687 images were used to compare non-inpainted methods to inpainted methods for CAL predictions. Statistical analyses were conducted using mean bias error (MBE), mean absolute error (MAE) and Dunn’s pairwise test comparing CAL at p=0.05. Comparator p-values demonstrated statistically significant improvement in CAL prediction accuracy between corresponding inpainted and non-inpainted methods with a MAE of 1.04mm and 1.50mm respectively. The Dunn’s pairwise test indicated a statistically significant improvement in CAL prediction accuracy between both inpainted methods compared to their non-inpainted counterparts, with the best performing methods achieving a Dunn’s pairwise value of -63.89. This study demonstrates the superiority of using a generative adversarial inpainting network with partial convolutions to predict CAL from bitewing images.

Cast suppression in radiographs by generative adversarial networks

2021 ◽  
Author(s):  
Franko Hržić ◽  
Ivana Žužić ◽  
Sebastian Tschauner ◽  
Ivan Štajduhar
Keyword(s):  
Software Tool ◽  
Generative Adversarial Networks ◽  
X Rays ◽  
Generative Adversarial Network ◽  
Adversarial Network ◽  
Adversarial Networks ◽  
The Mean ◽  
Radiology Image

Abstract Injured extremities commonly need to be immobilized by casts to allow proper healing. We propose a method to suppress cast superimpositions in pediatric wrist radiographs based on the cycle generative adversarial network (CycleGAN) model. We retrospectively reviewed unpaired pediatric wrist radiographs (n = 9672) and sampled them into 2 equal groups, with and without cast. The test subset consisted of 718 radiographs with cast. We evaluated different quadratic input sizes (256, 512, and 1024 pixels) for U-Net and ResNet-based CycleGAN architectures in cast suppression, quantitatively and qualitatively. The mean age was 11 ± 3 years in images containing cast (n = 4836), and 11 ± 4 years in castless samples (n = 4836). A total of 5956 X-rays had been done in males and 3716 in females. A U-Net 512 CycleGAN performed best (P ≤ .001). CycleGAN models successfully suppressed casts in pediatric wrist radiographs, allowing the development of a related software tool for radiology image viewers.

scholarly journals A Comparison of Models for Estimating Solar Radiation from Sunshine Duration in Croatia

2020 ◽  
Vol 2020 ◽  
pp. 1-14
Author(s):  
Tihomir Betti ◽  
Ivana Zulim ◽  
Slavica Brkić ◽  
Blanka Tuka
Keyword(s):  
Mean Absolute Error ◽  
Sunshine Duration ◽  
Model Performance ◽  
Absolute Error ◽  
Mean Bias Error ◽  
Percentage Error ◽  
Bias Error ◽  
Mean Square ◽  
Using Data ◽  
Statistical Indicators

The performance of seventeen sunshine-duration-based models has been assessed using data from seven meteorological stations in Croatia. Conventional statistical indicators are used as numerical indicators of the model performance: mean absolute percentage error (MAPE), mean bias error (MBE), mean absolute error (MAE), and root-mean-square error (RMSE). The ranking of the models was done using the combination of all these parameters, all having equal weights. The Rietveld model was found to perform the best overall, followed by Soler and Dogniaux-Lemoine monthly dependent models. For three best-performing models, new adjusted coefficients are calculated, and they are validated using separate dataset. Only the Dogniaux-Lemoine model performed better with adjusted coefficients, but across all analysed locations, the adjusted models showed improvement in reduced maximum percentage error.

scholarly journals Least kth-Order and Rényi Generative Adversarial Networks

2021 ◽  
pp. 1-38
Author(s):  
Himesh Bhatia ◽  
William Paul ◽  
Fady Alajaji ◽  
Bahman Gharesifard ◽  
Philippe Burlina
Keyword(s):  
Loss Function ◽  
Degrees Of Freedom ◽  
Natural Extension ◽  
Absolute Error ◽  
Loss Functions ◽  
Generative Adversarial Networks ◽  
Data Sets ◽  
Adversarial Networks ◽  
Information Theoretic Measures ◽  
Jensen Shannon Divergence

Abstract We investigate the use of parameterized families of information-theoretic measures to generalize the loss functions of generative adversarial networks (GANs) with the objective of improving performance. A new generator loss function, least kth-order GAN (LkGAN), is introduced, generalizing the least squares GANs (LSGANs) by using a kth-order absolute error distortion measure with k≥1 (which recovers the LSGAN loss function when k=2). It is shown that minimizing this generalized loss function under an (unconstrained) optimal discriminator is equivalent to minimizing the kth-order Pearson-Vajda divergence. Another novel GAN generator loss function is next proposed in terms of Rényi cross-entropy functionals with order α>0, α≠1. It is demonstrated that this Rényi-centric generalized loss function, which provably reduces to the original GAN loss function as α→1, preserves the equilibrium point satisfied by the original GAN based on the Jensen-Rényi divergence, a natural extension of the Jensen-Shannon divergence. Experimental results indicate that the proposed loss functions, applied to the MNIST and CelebA data sets, under both DCGAN and StyleGAN architectures, confer performance benefits by virtue of the extra degrees of freedom provided by the parameters k and α, respectively. More specifically, experiments show improvements with regard to the quality of the generated images as measured by the Fréchet inception distance score and training stability. While it was applied to GANs in this study, the proposed approach is generic and can be used in other applications of information theory to deep learning, for example, the issues of fairness or privacy in artificial intelligence.

scholarly journals ELECTRICAL PARAMETERS ESTIMATION OF SINGLE DIODE PV MODULE MODEL USING HYBRID METAHEURISTIC ALGORIHM

2022 ◽  
Vol 26 (1) ◽  
pp. 64-78
Author(s):  
Mawj M. Abbas ◽  
◽  
Dhiaa H. Muhsen ◽  
Keyword(s):  
Differential Evolution ◽  
Execution Time ◽  
Absolute Error ◽  
Standard Test ◽  
Parameters Estimation ◽  
Coefficient Of Determination ◽  
Mean Bias Error ◽  
Bias Error ◽  
De Algorithm ◽  
Pv Module

In this paper, an improved hybrid algorithm called differential evolution with integrated mutation per iteration (DEIM) is proposed to extract five parameters of single-diode PV module model obtained by combining differential evolution (DE) algorithm and electromagnetic-like (EML) algorithm. The EML algorithm's attraction-repulsion idea is employed in DEIM in order to enhance the mutation process of DE. The proposed algorithm is validated with other methods using experimental I-V data. The results of presented method reveal that simulated I-V characteristics have a high degree of agreement with experimental ones. The proposed model has an average root mean square error of 0.062A, an absolute error of 0.0452A, a mean bias error of 0.006A, a coefficient of determination of 0.992, a standard test deviation around 0.04540, and 15.33sec as execution time. The results demonstrate that the proposed method is better in terms of the accuracy and execution time (convergence) when compared with other methods where provide less errors.

scholarly journals Validity and Absolute Reliability of Axial Vertebral Rotation Measurements in Thoracic and Lumbar Vertebrae

2021 ◽  
Vol 11 (23) ◽  
pp. 11084
Author(s):  
José Hurtado-Avilés ◽  
Vicente J. León-Muñoz ◽  
Pilar Andújar-Ortuño ◽  
Fernando Santonja-Renedo ◽  
Mónica Collazo-Diéguez ◽  
...  
Keyword(s):  
Lumbar Vertebrae ◽  
Vertebral Rotation ◽  
Mean Bias Error ◽  
Bias Error ◽  
Thoracic Vertebrae ◽  
X Rays ◽  
Validity And Reliability ◽  
Axial Vertebral Rotation ◽  
Manual Methods ◽  
Absolute Reliability

Axial vertebral rotation (AVR) and Cobb angles are the essential parameters to analyse different types of scoliosis, including adolescent idiopathic scoliosis. The literature shows significant discrepancies in the validity and reliability of AVR measurements taken in radiographic examinations, according to the type of vertebra. This study’s scope evaluated the validity and absolute reliability of thoracic and lumbar vertebrae AVR measurements, using a validated software based on Raimondi’s method in digital X-rays that allowed measurement with minor error when compared with other traditional, manual methods. Twelve independent evaluators measured AVR on the 74 most rotated vertebrae in 42 X-rays with the software on three separate occasions, with one-month intervals. We have obtained a gold standard for the AVR of vertebrae. The validity and reliability of the measurements of the thoracic and lumbar vertebrae were studied separately. Measurements that were performed on lumbar vertebrae were shown to be 3.6 times more valid than those performed on thoracic, and with almost an equal reliability (1.38° ± 1.88° compared to −0.38° ± 1.83°). We can conclude that AVR measurements of the thoracic vertebrae show a more significant Mean Bias Error and a very similar reliability than those of the lumbar vertebrae.

Weakly supervised pneumonia localization in chest x‐rays using generative adversarial networks

2021 ◽  
Author(s):  
Krishna Nand Keshavamurthy ◽  
Carsten Eickhoff ◽  
Krishna Juluru
Keyword(s):  
Generative Adversarial Networks ◽  
X Rays ◽  
Adversarial Networks ◽  
Weakly Supervised

scholarly journals Data Augmentation Using Generative Adversarial Networks (GANs) For GAN-Based Detection Of Pneumonia And COVID-19 In Chest X-Ray Images

2021 ◽  
Author(s):  
Saman Motamed ◽  
Patrik Rogalla ◽  
Farzad Khalvati
Keyword(s):  
Neural Networks ◽  
Data Augmentation ◽  
Generative Models ◽  
Training Data ◽  
Generative Adversarial Networks ◽  
X Rays ◽  
X Ray ◽  
Standard Data ◽  
Adversarial Networks ◽  
Augmentation Techniques

Abstract Successful training of convolutional neural networks (CNNs) requires a substantial amount of data. With small datasets networks generalize poorly. Data Augmentation techniques improve the generalizability of neural networks by using existing training data more effectively. Standard data augmentation methods, however, produce limited plausible alternative data. Generative Adversarial Networks (GANs) have been utilized to generate new data and improve the performance of CNNs. Nevertheless, data augmentation techniques for training GANs are under-explored compared to CNNs. In this work, we propose a new GAN architecture for augmentation of chest X-rays for semi-supervised detection of pneumonia and COVID-19 using generative models. We show that the proposed GAN can be used to effectively augment data and improve classification accuracy of disease in chest X-rays for pneumonia and COVID-19. We compare our augmentation GAN model with Deep Convolutional GAN and traditional augmentation methods (rotate, zoom, etc) on two different X-ray datasets and show our GAN-based augmentation method surpasses other augmentation methods for training a GAN in detecting anomalies in X-ray images.

scholarly journals Experimental and Theoretical Study on the Internal Convective and Radiative Heat Transfer Coefficients for a Vertical Wall in a Residential Building

Energies ◽  
2021 ◽  
Vol 14 (18) ◽  
pp. 5953
Author(s):  
Piotr Michalak
Keyword(s):  
Heat Transfer ◽  
Vertical Wall ◽  
Experimental Studies ◽  
Residential Building ◽  
Absolute Error ◽  
Mean Bias Error ◽  
Bias Error ◽  
Single Family ◽  
Transfer Coefficients ◽  
Heat Transfer Coefficients

Experimental studies on internal convective (CHTC) and radiative (RHTC) heat transfer coefficients are very rarely conducted in real conditions during the normal use of buildings. This study presents the results of measurements of CHTC and RHTC for a vertical wall, taken in a selected room of a single-family building during its everyday use. Measurements were performed using HFP01 heat flux plates, Pt1000 sensors for internal air and wall surface temperatures and a globe thermometer for mean radiant temperature measured in 10 min intervals. Measured average CHTC and RHTC amounted to 1.15 W/m2K and 5.45 W/m2K, compared to the 2.50 W/m2K and 5.42 W/m2K recommended by the EN ISO 6946, respectively. To compare with calculated CHTC, 14 correlations based on the temperature difference were applied. Obtained values were from 1.31 W/m2K (given by Min et al.) to 3.33 W/m2K (Wilkes and Peterson), and in all cases were greater than the 1.15 W/m2K from measurements. The average value from all models amounted to 2.02 W/m2K, and was greater than measurements by 75.6%. The quality of models was also estimated using average absolute error (AAE), average biased error (ABE), mean absolute error (MAE) and mean bias error (MBE). Based on these techniques, the model of Fohanno and Polidori was identified as the best with AAE = 68%, ABE = 52%, MAE = 0.41 W/m2K and MBE = 0.12 W/m2K.

Sign in / Sign up

Export Citation Format

Share Document