Classification Uncertainty Research Articles

A New Belief-Based Bidirectional Transfer Classification Method.

In pattern classification, we may have a few labeled data points in the target domain, but a number of labeled samples are available in another related domain (called the source domain). Transfer learning can solve such classification problems via the knowledge transfer from source to target domains. The source and target domains can be represented by heterogeneous features. There may exist uncertainty in domain transformation, and such uncertainty is not good for classification. The effective management of uncertainty is important for improving classification accuracy. So, a new belief-based bidirectional transfer classification (BDTC) method is proposed. In BDTC, the intraclass transformation matrix is estimated at first for mapping the patterns from source to target domains, and this matrix can be learned using the labeled patterns of the same class represented by heterogeneous domains (features). The labeled patterns in the source domain are transferred to the target domain by the corresponding transformation matrix. Then, we learn a classifier using all the labeled patterns in the target domain to classify the objects. In order to take full advantage of the complementary knowledge of different domains, we transfer the query patterns from target to source domains using the K-NN technique and do the classification task in the source domain. Thus, two pieces of classification results can be obtained for each query pattern in the source and target domains, but the classification results may have different reliabilities/weights. A weighted combination rule is developed to combine the two classification results based on the belief functions theory, which is an expert at dealing with uncertain information. We can efficiently reduce the uncertainty of transfer classification via the combination strategy. Experiments on some domain adaptation benchmarks show that our method can effectively improve classification accuracy compared with other related methods.

A Superpixel Spatial Intuitionistic Fuzzy C-Means Clustering Algorithm for Unsupervised Classification of High Spatial Resolution Remote Sensing Images

This paper proposes a superpixel spatial intuitionistic fuzzy C-means (SSIFCM) clustering algorithm to address the problems of misclassification, salt and pepper noise, and classification uncertainty arising in the pixel-level unsupervised classification of high spatial resolution remote sensing (HSRRS) images. To reduce information redundancy and ensure noise immunity and image detail preservation, we first use a superpixel segmentation to obtain the local spatial information of the HSRRS image. Secondly, based on the bias-corrected fuzzy C-means (BCFCM) clustering algorithm, the superpixel spatial intuitionistic fuzzy membership matrix is constructed by counting an intuitionistic fuzzy set and spatial function. Finally, to minimize the classification uncertainty, the local relation between adjacent superpixels is used to obtain the classification results according to the spectral features of superpixels. Four HSRRS images of different scenes in the aerial image dataset (AID) are selected to analyze the classification performance, and fifteen main existing unsupervised classification algorithms are used to make inter-comparisons with the proposed SSIFCM algorithm. The results show that the overall accuracy and Kappa coefficients obtained by the proposed SSIFCM algorithm are the best within the inter-comparison of fifteen algorithms, which indicates that the SSIFCM algorithm can effectively improve the classification accuracy of HSRRS image.

Detailed investigation of discrepancies in Köppen-Geiger climate classification using seven global gridded products

The Köppen-Geiger (KG) climate classification has been widely used to determine the climate at global and regional scales using precipitation and temperature data. KG maps are typically developed using a single product; however, uncertainties in KG climate types resulting from different precipitation and temperature datasets have not been explored in detail. Here, we assess seven global datasets to show uncertainties in KG classification from 1980 to 2017. Using a pairwise comparison at global and zonal scales, we quantify the similarity among the seven KG maps. Gauge- and reanalysis-based KG maps have a notable difference. Spatially, the highest and lowest similarity is observed for the North and South Temperate zones, respectively. Notably, 17% of grids among the seven maps show variations even in the major KG climate types, while 35% of grids are described by more than one KG climate subtype. Strong uncertainty is observed in south Asia, central and south Africa, western America, and northeastern Australia. We created two KG master maps (0.5° resolution) by merging the climate maps directly and by combining the precipitation and temperature data from the seven datasets. These master maps are more robust than the individual ones showing coherent spatial patterns. This study reveals the large uncertainty in climate classification and offers two robust KG maps that may help to better evaluate historical climate and quantify future climate shifts.

Incorporating Measurement Error in Astronomical Object Classification

Most general-purpose classification methods, such as support-vector machine (SVM) and random forest (RF), fail to account for an unusual characteristic of astronomical data: known measurement error uncertainties. In astronomical data, this information is often given in the data but discarded because popular machine learning classifiers cannot incorporate it. We propose a simulation-based approach that incorporates heteroscedastic measurement error into an existing classification method to better quantify uncertainty in classification. The proposed method first simulates perturbed realizations of the data from a Bayesian posterior predictive distribution of a Gaussian measurement error model. Then, a chosen classifier is fit to each simulation. The variation across the simulations naturally reflects the uncertainty propagated from the measurement errors in both labeled and unlabeled data sets. We demonstrate the use of this approach via two numerical studies. The first is a thorough simulation study applying the proposed procedure to SVM and RF, which are well-known hard and soft classifiers, respectively. The second study is a realistic classification problem of identifying high-z (2.9 ≤ z ≤ 5.1) quasar candidates from photometric data. The data are from merged catalogs of the Sloan Digital Sky Survey, the Spitzer IRAC Equatorial Survey, and the Spitzer-HETDEX Exploratory Large-Area Survey. The proposed approach reveals that out of 11,847 high-z quasar candidates identified by a random forest without incorporating measurement error, 3146 are potential misclassifications with measurement error. Additionally, out of 1.85 million objects not identified as high-z quasars without measurement error, 936 can be considered new candidates with measurement error.

Uncertainty in lung cancer stage for survival estimation via set‐valued classification

The difficulty in identifying cancer stage in health care claims data has limited oncology quality of care and health outcomes research. We fit prediction algorithms for classifying lung cancer stage into three classes (stages I/II, stage III, and stage IV) using claims data, and then demonstrate a method for incorporating the classification uncertainty in survival estimation. Leveraging set‐valued classification and split conformal inference, we show how a fixed algorithm developed in one cohort of data may be deployed in another, while rigorously accounting for uncertainty from the initial classification step. We demonstrate this process using SEER cancer registry data linked with Medicare claims data.

Revision of 18 ichneumonid fossil species (Hymenoptera, Ichneumonidae) highlights the need for open nomenclature in palaeontology

The fossil record of Darwin wasps (Ichneumonidae) is extremely understudied relative to their enormous extant diversity, with only around 300 species described. Moreover, the taxonomic placement of many of the fossils is based on an outdated classification system. We here revise 18 ichneumonid fossils, all described before the most comprehensive revision of ichneumonid classification by Henry Townes. After a careful reinterpretation of character evidence, we leave the original placement of only five fossils, while expressing uncertainty about the placement of two of them, Pimpla ? seyrigi Theobald and Polysphincta ? inundata Brues, by following the principles of open nomenclature. In addition, we move Parapimpla rhenana Theobald, 1937 stat. rev. from Pimplinae to Ctenopelmatinae. We describe a new phygadeuontine genus to accommodate two fossils previously described in Pimplinae, Armadilleon morticinus (Brues) gen. et comb. nov. and A. petrorum (Brues) gen. et comb. nov. Finally, we change the genus and, almost always, the subfamily placement for ten fossils: Hallocinetus ? arvernus (Piton) comb. nov., Dimophora ? longicornis (Theobald) comb. nov., Dimophora ? wickhami (Cockerell) comb. nov., Lycorina ? indura (Theobald) comb. nov., Acerataspis ? revelata (Brues) comb. nov., Hypsicera ? solidata (Brues) comb. nov., Orthocentrus ? mortuaria (Brues) comb. nov., Zagryphus tilloyi (Theobald) comb. nov., Lithoserix antiquus (Saussure) comb. nov. and Monoblastus ? senilis (Brues) comb. nov. Our revision highlights the need for the re-interpretation of perhaps a majority of ichneumonid fossils and for widely adopting the open nomenclature framework. This framework allows uncertainty in fossil classification to be expressed in an intuitive and explicit manner, which contributes to alleviating misinterpretation of the palaeontological literature, for instance in phylogenetic dating studies.

Automatic recognition of flow cytometric phytoplankton functional groups using convolutional neural networks

AbstractThe variability of phytoplankton distribution has been unraveled by high‐frequency measurements. Such a resolution can be approached by automated pulse‐shape recording flow cytometry (AFCM) operating at hourly sampling resolution. AFCM records morphological and physiological traits as single‐cell optical pulse shapes that can be used to classify cells into phytoplankton functional groups (PFGs). However, the associated manual post‐processing of the data coupled with the increasing size and number of datasets is time‐consuming and error‐prone. Machine learning models are increasingly used to run automatic classification. Yet, most of the existing methods either present a long training process, need to manually design features from the raw optical pulse shapes, or are dedicated to images only. In this study, we present a convolutional neural network (CNN) to classify several PFGs using AFCM pulse shapes. The uncertainties of manual classification were first estimated by comparing experts' recognition of six PFGs. Consensual particles from the manual PFG classification were used to train and validate the CNN. The CNN obtained competitive performances compared to other models used in the literature and remained robust across several sampling areas, and instrumental hardware and settings. Finally, we assessed the ability of this classifier to predict phytoplankton counts at a Mediterranean coastal station and from a cruise in the South‐West Indian Ocean, providing a comparison with the manual classification over 3‐month periods and a 2h frequency. These promising results strengthen the near real‐time observation of PFGs, especially required with the increasing use of AFCM in monitoring research programs.

Intuitionistic Fuzzy-Based Three-Way Label Enhancement for Multi-Label Classification

Multi-label classification deals with the determination of instance-label associations for unseen instances. Although many margin-based approaches are delicately developed, the uncertainty classifications for those with smaller separation margins remain unsolved. The intuitionistic fuzzy set is an effective tool to characterize the concept of uncertainty, yet it has not been examined for multi-label cases. This paper proposed a novel model called intuitionistic fuzzy three-way label enhancement (IFTWLE) for multi-label classification. The IFTWLE combines label enhancement with an intuitionistic fuzzy set under the framework of three-way decisions. For unseen instances, we generated the pseudo-label for label uncertainty evaluation from a logical label-based model. An intuitionistic fuzzy set-based instance selection principle seamlessly bridges logical label learning and numerical label learning. The principle is hierarchically developed. At the label level, membership and non-membership functions are pair-wisely defined to measure the local uncertainty and generate candidate uncertain instances. After upgrading to the instance level, we select instances from the candidates for label enhancement, whereas they remained unchanged for the remaining. To the best of our knowledge, this is the first attempt to combine logical label learning with numerical label learning into a unified framework for minimizing classification uncertainty. Extensive experiments demonstrate that, with the selectively reconstructed label importance, IFTWLE achieves statistically superior over the state-of-the-art multi-label classification algorithms in terms of classification accuracy. The computational complexity of this algorithm is On2mk, where n, m, and k denote the unseen instances count, label count, and average label-specific feature size, respectively.

Application of Improved MFDFA and D-S Evidence Theory in Fault Diagnosis

To improve the accuracy of centrifugal pump fault diagnosis, a novel fault diagnosis method based on improved multiple fractal detrended fluctuation analysis (MFDFA), the fusion of multi-sensing information derived from the back propagation (BP) neural network and the Dempster–Shafter (D-S) evidence theory, is accordingly proposed. Firstly, the multifractal spectral parameters of four sensor signals under four different operating conditions were extracted as centrifugal pump fault feature vectors using improved MFDFA and input to the BP neural network. Then, the basic trust assignment function was constructed by calculating trustworthiness (both local and global) as priori information, which is based on the output results of the neural networks specific to of each group of sensors. Finally, the basic trust assignment function was fused with decision processing in accordance with the D-S evidence combination rule in order to effectively achieve the multi-sensor information fusion centrifugal pump fault diagnosis. The experimental results show the multiple fractal spectrum feature parameters extracted by the improved MFDFA can accurately reflect the signal essence, and can be used as the fault feature vector. On this basis, this multi-sensor fault diagnosis reduces the uncertainty of fault classification and demonstrates improved accuracy compared to the single-sensor fault diagnosis thanks to being based on a combination of the BP neural networks and D-S evidence theory. Thereby, this method can facilitate accurate diagnosis of the centrifugal pump fault type with high confidence, subsequently providing a novel and alternative method to existing methods of diagnosis.

SAR Image Fusion Classification Based on the Decision-Level Combination of Multi-Band Information

Synthetic aperture radar (SAR) is an active coherent microwave remote sensing system. SAR systems working in different bands have different imaging results for the same area, resulting in different advantages and limitations for SAR image classification. Therefore, to synthesize the classification information of SAR images into different bands, an SAR image fusion classification method based on the decision-level combination of multi-band information is proposed in this paper. Within the proposed method, the idea of Dempster–Shafer evidence theory is introduced to model the uncertainty of the classification result of each pixel and used to combine the classification results of multiple band SAR images. The convolutional neural network is used to classify single-band SAR images. Calculate the belief entropy of each pixel to measure the uncertainty of single-band classification, and generate the basic probability assignment function. The idea of the term frequency-inverse document frequency in natural language processing is combined with the conflict coefficient to obtain the weight of different bands. Meanwhile, the neighborhood classification of each pixel in different band sensors is considered to obtain the total weight of each band sensor, generate weighted average BPA, and obtain the final ground object classification result after fusion. The validity of the proposed method is verified in two groups of multi-band SAR image classification experiments, and the proposed method has effectively improved the accuracy compared to the modified average approach.

Limit state Kriging modeling for reliability-based design optimization through classification uncertainty quantification

Reliability-based design optimization (RBDO) plays a vital role in considering the effect of uncertainties in the optimal design variables on the production reliability. Kriging-assisted RBDO methods can reduce the computational cost of conventional RBDO methods by replacing the time-consuming performance functions with Kriging models. Existing Kriging-assisted RBDO methods, however, are easy to fall into the low modeling efficiency issue or unsatisfied modeling accuracy issue because of the low utilization rate of sample resources. In this paper, an adaptive Kriging sampling strategy based on the Classification Uncertainty Quantification (KCUQ) was proposed. In KCUQ, the classification uncertainty of the Kriging model is sufficiently considered by (1) determining the new sample point based on the quantified misclassification probability and (2) checking the modeling accuracy based on the quantified number of misclassified random points. Moreover, KCUQ only updates the performance function with the largest classification error in each iteration such that all performance functions can be adaptively modeled based on their unique features. Two numerical case studies, vehicle side impact crashworthiness problem and the axle bridge turning parameters optimization application are used to demonstrate the performance of the proposed KCUQ method.

Quantifying uncertainty in machine learning classifiers for medical imaging.

Machine learning (ML) models in medical imaging (MI) can be of great value in computer aided diagnostic systems, but little attention is given to the confidence (alternatively, uncertainty) of such models, which may have significant clinical implications. This paper applied, validated, and explored a technique for assessing uncertainty in convolutional neural networks (CNNs) in the context of MI. We used two publicly accessible imaging datasets: a chest x-ray dataset (pneumonia vs. control) and a skin cancer imaging dataset (malignant vs. benign) to explore the proposed measure of uncertainty based on experiments with different class imbalance-sample sizes, and experiments with images close to the classification boundary. We also further verified our hypothesis by examining the relationship with other performance metrics and cross-checking CNN predictions and confidence scores with an expert radiologist (available in the Supplementary Information). Additionally, bounds were derived on the uncertainty metric, and recommendations for interpretability were made. With respect to training set class imbalance for the pneumonia MI dataset, the uncertainty metric was minimized when both classes were nearly equal in size (regardless of training set size) and was approximately 17% smaller than the maximum uncertainty resulting from greater imbalance. We found that less-obvious test images (those closer to the classification boundary) produced higher classification uncertainty, about 10-15 times greater than images further from the boundary. Relevant MI performance metrics like accuracy, sensitivity, and sensibility showed seemingly negative linear correlations, though none were statistically significant (p [Formula: see text] 0.05). The expert radiologist and CNN expressed agreement on a small sample of test images, though this finding is only preliminary. This paper demonstrated the importance of uncertainty reporting alongside predictions in medical imaging. Results demonstrate considerable potential from automatically assessing classifier reliability on each prediction with the proposed uncertainty metric.

Intelligent fault diagnosis based on sample weighted joint adversarial network

In recent years, adversarial-based deep domain adaptation (DDA) has attracted increasing attention in transferable fault diagnosis. However, most of the existing methods mainly focus on the elimination of global distribution discrepancies, and neglect the exploration of category-specific distribution characteristics, resulting in unsatisfactory diagnostic results in complex scenarios. To avoid this drawback, a novel sample weighted joint adversarial network (SWJAN) is proposed in this paper, which exploits the category information to enhance the joint domain adaptability of adversarial learning. Specifically, in SWJAN, the generated feature vector for domain recognition is decomposed into a feature matrix according to the label probability, so that the category distribution of the feature space can be captured and aligned. As a result, not only the traditional domain-wise adaptation, but also the class-wise feature matching can be emphasized simultaneously. Moreover, since it is risky to adapt the distribution on low-confidence data, the weight of each sample is dynamically adjusted according to the classification uncertainty, so as to control the influence of ambiguous data on adversarial learning and suppress negative transfer near the decision boundary. Experimental analysis, including cross-domain fault diagnosis of the gearbox and rolling bearing, verifies the practicability and superiority of SWJAN in engineering applications.

Star–Galaxy Image Separation with Computationally Efficient Gaussian Process Classification

Abstract We introduce a novel method for discerning optical telescope images of stars from those of galaxies using Gaussian processes (GPs). Although applications of GPs often struggle in high-dimensional data modalities such as optical image classification, we show that a low-dimensional embedding of images into a metric space defined by the principal components of the data suffices to produce high-quality predictions from real large-scale survey data. We develop a novel method of GP classification hyperparameter training that scales approximately linearly in the number of image observations, which allows for application of GP models to large-size Hyper Suprime-Cam Subaru Strategic Program data. In our experiments, we evaluate the performance of a principal component analysis embedded GP predictive model against other machine-learning algorithms, including a convolutional neural network and an image photometric morphology discriminator. Our analysis shows that our methods compare favorably with current methods in optical image classification while producing posterior distributions from the GP regression that can be used to quantify object classification uncertainty. We further describe how classification uncertainty can be used to efficiently parse large-scale survey imaging data to produce high-confidence object catalogs.

A proposed artificial intelligence workflow to address application challenges leveraged on algorithm uncertainty

SummaryArtificial Intelligence (AI) has achieved state-of-the-art performance in medical imaging. However, most algorithms focused exclusively on improving the accuracy of classification while neglecting the major challenges in a real-world application. The opacity of algorithms prevents users from knowing when the algorithms might fail. And the natural gap between training datasets and the in-reality data may lead to unexpected AI system malfunction. Knowing the underlying uncertainty is essential for improving system reliability. Therefore, we developed a COVID-19 AI system, utilizing a Bayesian neural network to calculate uncertainties in classification and reliability intervals of datasets. Validated with four multi-region datasets simulating different scenarios, our approach was proved to be effective to suggest the system failing possibility and give the decision power to human experts in time. Leveraging on the complementary strengths of AI and health professionals, our present method has the potential to improve the practicability of AI systems in clinical application.

Assessment of uncertainty and credibility of predictions by the OECD QSAR Toolbox automated read-across workflow for predicting acute oral toxicity

The platform of OECD Toolbox version 4.5 was used for building an automated decision tree for filling data gaps for rat acute oral toxicity (AOT) by read-across (RA). Our previous publications have described the workflow of the AOT tree and conducted verification and validation studies on it. The overall uncertainty in the AOT workflow is low as the similarity in mechanistic probability, metabolism and 2D structure are maximized in the RA analogue selection process. The endpoint, rat oral LD50, is well-defined and has universal regulatory acceptance. Since OECD test guidelines are followed in generating the database, the data are widely recognized to be of the highest quality. The credibility of the workflow is high as it meets the critical factors of being based on confirmed assumptions, having demonstrated concordance and consistency, permitting the ability to explain AOT-related mechanisms and modes of action, and being simple in design. Additionally, the Z-score and probability distribution methods of assessing the uncertainty of a particular RA are discussed. Two examples of numerical and classification uncertainty are presented. These cases represent the extremes observed in a series of target chemical-based predictions that the authors observed when testing the workflow. The reliability and relevance associated with the workflow are high. However, the completeness and weights-of-evidence varied markedly among possible RA scenarios and particular target substances.

Effect of passive metallic layers on muon energy estimation by means of deflection angle for muon scattering tomography: a comparative study based on GEANT4 simulations

In the tomographic configurations based on the muon scattering, the angular variation with respect to the kinetic energy indirectly brings forth the ability to coarsely predict the kinetic energy by using the deflection angle owing to the detector layers. Nevertheless, the angular deviation due to the detector components is expected to be minuscule in addition to a relatively high uncertainty in the case of the plastic scintillators. In the present study, we contrast our current tomographic prototype, which consists of the detector layers manufactured from polyvinyl toluene besides a detector accuracy of 1 mrad, with an alternative hodoscope scheme containing stainless steel layers by aiming to investigate the three-group energy structure. Initially, we determine the average deflection angles together with the corresponding standard deviations for our present setup as well as for the alternative scheme by means of the GEANT4 simulations. In the second place, we express a brace of misclassification probabilities founded on the standard deviations where the first procedure assumes a linear finite approximation, whereas the latter approach rests on a positively defined modified Gaussian distribution. Upon our simulation results, we demonstrate that the introduced stainless steel layers in the proposed hodoscope setup do not only serve to augment the average deflection angles, but they also diminish the misclassification probabilities, therewith reducing the classification uncertainty apart from an improved detection performance.

Seeking new physics in cosmology with Bayesian neural networks: Dark energy and modified gravity

We study the potential of Bayesian neural networks (BNNs) to detect new physics in the dark matter power spectrum, concentrating here on evolving dark energy and modifications to general relativity. After introducing a new technique to quantify classification uncertainty in BNNs, we train two BNNs on mock matter power spectra produced using the publicly available code react in the $k$ range $(0.01\ensuremath{-}2.5)\text{ }\text{ }h{\mathrm{Mpc}}^{\ensuremath{-}1}$ and redshift bins (0.1, 0.478, 0.783, 1.5) with Euclid-like noise. The first network classifies spectra into five labels including $\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$, $f(R)$, $w\mathrm{CDM}$, Dvali-Gabadadze-Porrati gravity and a ``random'' class, whereas the second is trained to distinguish $\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$ from non-$\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$. Both networks achieve a comparable training, validation and test accuracy of $\ensuremath{\sim}95%$. Each network is also capable of correctly classifying spectra with deviations from $\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$ that were not included in the training set, demonstrated with spectra generated using the growth index $\ensuremath{\gamma}$. To obtain an indication of the BNNs classification capability, we compute the smallest deviation from $\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$ such that the noise-averaged non-$\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$ classification probability is at least 95% according to our estimated error quantification, finding these bounds to be ${f}_{R0}\ensuremath{\lesssim}{10}^{\ensuremath{-}7}$, ${\mathrm{\ensuremath{\Omega}}}_{\mathrm{rc}}\ensuremath{\lesssim}{10}^{\ensuremath{-}2}$, $\ensuremath{-}1.05\ensuremath{\lesssim}{w}_{0}\ensuremath{\lesssim}0.95$, $\ensuremath{-}0.2\ensuremath{\lesssim}{w}_{a}\ensuremath{\lesssim}0.2$, and $0.52\ensuremath{\lesssim}\ensuremath{\gamma}\ensuremath{\lesssim}0.59$. The bounds on $f(R)$ can be improved by training a specialist network to distinguish solely between $\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$ and $f(R)$ power spectra which can detect a nonzero ${f}_{R0}$ at $\mathcal{O}({10}^{\ensuremath{-}8})$. We expect that further developments, such as the inclusion of smaller length scales or additional extensions to $\mathrm{\ensuremath{\Lambda}}\mathrm{CDM}$, will only improve the potential of BNNs to indicate the presence of new physics in cosmological datasets, regardless of the underlying theory.

Uncertainty meets 3D-spatial feature in colonoscopic polyp-size determination

ABSTRACT This paper presents a new 3D-spatial feature extraction from a 2D colonoscopic image for polyp-size estimation. The polyp-size estimation poses potential demands on colonoscopy, since an endoscopist’s subjective estimation is apt to result in uncertain polyp-size determination. This uncertain determination is derived from the lack of 3D-spatial information. Even though a previous work clarified that depth estimation mitigates uncertainty in binary polyp-size classification, precise estimation of 3D structure from a colonoscopic image(s) remains an unsolved challenge in medical image analysis. This work proposes a 3D-spatial feature that expresses a polyp’s precise 3D shape to mitigate uncertainty in polyp-size determination. First, we introduce an accurate depth estimation method to capture the 3D structure of a colon. Next, we integrate depth estimation and polyp localisation to extract a 3D polyp shape as a feature. Finally, we achieve polyp-size estimation by statistical learning of extracted features. The experimental results demonstrated the validity both of our depth estimation and 3D-spatial feature. Compared with deep RGB and RGB-D convolutional neural networks (CNNs), a shallow CNN with the proposed 3D-spatial feature achieved a more accurate polyp-size estimation with a mean absolute error of 1.36 mm, whereas the one of the deep CNN is 3.11 mm.

Developing a GIS-based rough fuzzy set granulation model to handle spatial uncertainty for hydrocarbon structure classification, case study: Fars domain, Iran

ABSTRACT It is well agreed that geologic risk occurs during hydrocarbon exploration because diverse uncertainties accompany the entire hydrocarbon system parameters such as the source rock, reservoir rock, trap and seal rock. In order to overcome such attributes with uncertainties, a number of soft computing methods are used. Information granules could be provided by the Rough Fuzzy Set Granulation (RFSG) with a thorough quality evaluation. This is capable of attribute reduction that has been claimed to be essential in investigating the hydrocarbon systems. This paper is an endeavor to recommend a Geospatial Information System (GIS)-based model with the aim of categorizing the hydrocarbon structures map consistent with the uncertainty range concepts of geologic risk in the rough fuzzy sets and granular computing. The model used the RFSG for the attribute reduction by a Decision Logic language (DL-language). The RFSG was employed in order to classify hydrocarbon structures according to geological risk and extract the fuzzy rules with a predefined range of uncertainty. In order to assess the precisions of the fuzzy decisions on the hydrocarbon structure classification, the fuzzy entropy and fuzzy cross-entropy are applied. The proposed RFSG model applied for 62 structures as the training data, average fuzzy entropy has been calculated as 0.85, whereas the average fuzzy cross-entropy has been calculated 0.18. As it can be discerned, just seven structures had cross-entropies greater than 0.1, while three structures were larger than 0.3. It is implied that the precision of the proposed model is about 89%. The results yielded two reductions for the condition attributes and 11 fuzzy rules being filtered by the granular computing values.