RESUMEN

Ultrasound images are susceptible to various forms of quality degradation that negatively impact diagnosis. Common degradations include speckle noise, Gaussian noise, salt and pepper noise, and blurring. This research proposes an accurate ultrasound image denoising strategy based on firstly detecting the noise type, then, suitable denoising methods can be applied for each corruption. The technique depends on convolutional neural networks to categorize the type of noise affecting an input ultrasound image. Pre-trained convolutional neural network models including GoogleNet, VGG-19, AlexNet and AlexNet-support vector machine (SVM) are developed and trained to perform this classification. A dataset of 782 numerically generated ultrasound images across different diseases and noise types is utilized for model training and evaluation. Results show AlexNet-SVM achieves the highest accuracy of 99.2% in classifying noise types. The results indicate that, the present technique is considered one of the top-performing models is then applied to real ultrasound images with different noise corruptions to demonstrate efficacy of the proposed detect-then-denoise system. RESEARCH HIGHLIGHTS: Proposes an accurate ultrasound image denoising strategy based on detecting noise type first. Uses pre-trained convolutional neural networks to categorize noise type in input images. Evaluates GoogleNet, VGG-19, AlexNet, and AlexNet-support vector machine (SVM) models on a dataset of 782 synthetic ultrasound images. AlexNet-SVM achieves highest accuracy of 99.2% in classifying noise types. Demonstrates efficacy of the proposed detect-then-denoise system on real ultrasound images.

RESUMEN

Introduction: Alzheimer's Disease (AD) is a degenerative brain disorder characterized by cognitive and memory dysfunctions. The early detection of AD is necessary to reduce the mortality rate through slowing down its progression. The prevention and detection of AD is the emerging research topic for many researchers. The structural Magnetic Resonance Imaging (sMRI) is an extensively used imaging technique in detection of AD, because it efficiently reflects the brain variations. Methods: Machine learning and deep learning models are widely applied on sMRI images for AD detection to accelerate the diagnosis process and to assist clinicians for timely treatment. In this article, an effective automated framework is implemented for early detection of AD. At first, the Region of Interest (RoI) is segmented from the acquired sMRI images by employing Otsu thresholding method with Tunicate Swarm Algorithm (TSA). The TSA finds the optimal segmentation threshold value for Otsu thresholding method. Then, the vectors are extracted from the RoI by applying Local Binary Pattern (LBP) and Local Directional Pattern variance (LDPv) descriptors. At last, the extracted vectors are passed to Deep Belief Networks (DBN) for image classification. Results and Discussion: The proposed framework achieves supreme classification accuracy of 99.80% and 99.92% on the Alzheimer's Disease Neuroimaging Initiative (ADNI) and Australian Imaging, Biomarker and Lifestyle flagship work of ageing (AIBL) datasets, which is higher than the conventional detection models.

RESUMEN

This work investigated the high-throughput classification performance of microscopic images of mesenchymal stem cells (MSCs) using a hyperspectral imaging-based separable convolutional neural network (CNN) (H-SCNN) model. Human bone marrow mesenchymal stem cells (hBMSCs) were cultured, and microscopic images were acquired using a fully automated microscope. Flow cytometry (FCT) was employed for functional classification. Subsequently, the H-SCNN model was established. The hyperspectral microscopic (HSM) images were created, and the spatial-spectral combined distance (SSCD) was employed to derive the spatial-spectral neighbors (SSNs) for each pixel in the training set to determine the optimal parameters. Then, a separable CNN (SCNN) was adopted instead of the classic convolutional layer. Additionally, cultured cells were seeded into 96-well plates, and high-functioning hBMSCs were screened using both manual visual inspection (MV group) and the H-SCNN model (H-SCNN group), with each group consisting of 96 samples. FCT served as the benchmark to compare the area under the curve (AUC), F1 score, accuracy (Acc), sensitivity (Sen), specificity (Spe), positive predictive value (PPV), and negative predictive value (NPV) between the manual and model groups. The best classification Acc was 0.862 when using window size of 9 and 12 SSNs. The classification Acc of the SCNN model, ResNet model, and VGGNet model gradually increased with the increase in sample size, reaching 89.56 ± 3.09, 80.61 ± 2.83, and 80.06 ± 3.01%, respectively at the sample size of 100. The corresponding training time for the SCNN model was significantly shorter at 21.32 ± 1.09 min compared to ResNet (36.09 ± 3.11 min) and VGGNet models (34.73 ± 3.72 min) (P < 0.05). Furthermore, the classification AUC, F1 score, Acc, Sen, Spe, PPV, and NPV were all higher in the H-SCNN group, with significantly less time required (P < 0.05). Microscopic images based on the H-SCNN model proved to be effective for the classification assessment of hBMSCs, demonstrating excellent performance in classification Acc and efficiency, enabling its potential to be a powerful tool in future MSCs research.

RESUMEN

BACKGROUND: In the agricultural sector, the early identification of plant diseases presents a pressing challenge. Throughout the growing season, plants remain vulnerable to an array of diseases. Failure to detect these diseases at their early stages can significantly compromise the overall yield, thereby reducing profitability for farmers. To address this issue, several researchers have introduced standard methods that leverage machine learning and deep learning techniques. However, many of these methods offer limited classification accuracy and often necessitate extensive training parameter adjustments. METHOD: The objective of this study is to develop a new deep learning-based technique for detecting and classifying plant diseases at earlier stages. Thus, this paper introduces a novel technique known as the deep belief network-based enhanced kernel extreme learning machine (DBN-EKELM) that identifies a disease automatically and performs effective classification. The initial phase involves data preprocessing to enhance quality of plant leaf images, facilitating the extraction of critical information. With the goal of achieving superior classification accuracy, this paper proposes the use of the DBN-EKELM technique for optimal plant leaf disease detection. Given that KELM parameters are highly sensitive to minor variations, proper parameter tuning is essential and introduces a novel binary gaining sharing knowledge-based optimization algorithm (NBGSK). RESULT: The efficacy of the proposed DBN-EKELM method is evaluated by comparing its performance with other conventional methods, considering various measures like accuracy, precision, specificity, sensitivity and F-measure. CONCLUSION: Experimental analyses demonstrate that the DBN-EKELM technique achieves an impressive rate of approximately 98.2%, 97%, 98.1%, 97.4% as well as 97.8%, surpassing other standard methods. © 2024 Society of Chemical Industry.

RESUMEN

It is shown that a more than significant amount of experimental data obtained in the field of studying systems based on thermosensitive hydrophilic polymers and reflected in the literature over the past decades makes the issue of their systematization and classification relevant. This, in turn, makes relevant the question of choosing the appropriate classification criteria. It is shown that the basic classification feature can be the number of phase transition stages, which can vary from one to four or more depending on the nature of the temperature-sensitive system. In this work, the method of inverse phase portraits is proposed for the first time. It was intended, among other things, to identify the number of phase transition stages. Moreover, the accuracy of this method significantly exceeds the accuracy of the previously used method of direct phase portraits since, for the first time, the operation of numerical differentiation is replaced by the operation of numerical integration. A specific example of the application of the proposed method for the analysis of a previously studied temperature-sensitive system is presented. It is shown that this method also allows for a quantitative comparison between the results obtained by the differential calorimetry method and the turbidimetry method. Issues related to increasing the resolution of the method of direct phase portraits are discussed.

RESUMEN

This study was designed to evaluate the classification accuracy of the Warrington's Recognition Memory Test (RMT) in 167 patients (97 or 58.1% men; MAge = 40.4; MEducation= 13.8) medically referred for neuropsychological evaluation against five psychometrically defined criterion groups. At the optimal cutoff (≤42), the RMT produced an acceptable combination of sensitivity (.36-.60) and specificity (.85-.95), correctly classifying 68.4-83.3% of the sample. Making the cutoff more conservative (≤41) improved specificity (.88-.95) at the expense of sensitivity (.30-.60). Lowering the cutoff to ≤40 achieved uniformly high specificity (.91-.95) but diminished sensitivity (.27-.48). RMT scores were unrelated to lateral dominance, education, or gender. The RMT was sensitive to a three-way classification of performance validity (Pass/Borderline/Fail), further demonstrating its discriminant power. Despite a notable decline in research studies focused on its classification accuracy within the last decade, the RMT remains an effective free-standing PVT that is robust to demographic variables. Relatively low sensitivity is its main liability. Further research is needed on its cross-cultural validity (sensitivity to limited English proficiency).

RESUMEN

Predicting repetitive transcranial magnetic stimulation (rTMS) treatment outcomes in major depressive disorder (MDD) could reduce the financial and psychological risks of treatment failure. We systematically reviewed and meta-analyzed studies that leveraged neurophysiological and neuroimaging markers to predict rTMS response in MDD. Five databases were searched from inception to May 25, 2023. The primary meta-analytic outcome was predictive accuracy pooled from classification models. Regression models were summarized qualitatively. A promising marker was identified if it showed a sensitivity and specificity of 80% or higher in at least two independent studies. Searching yielded 36 studies. Twenty-two classification modeling studies produced an estimated area under the summary receiver operating characteristic curve of 0.87 (95% CI = 0.83-0.92), with 86.8% sensitivity (95% CI = 80.6-91.2%) and 81.9% specificity (95% CI = 76.1-86.4%). Frontal theta cordance measured by electroencephalography is closest to proof of concept. Predicting rTMS response using neurophysiological and neuroimaging markers is promising for clinical decision-making. However, replications by different research groups are needed to establish rigorous markers.

Asunto(s)

RESUMEN

The accuracy of most classification methods is significantly affected by missing values. Therefore, this study aimed to propose a data imputation method to handle missing values through the application of nearest neighbor data and fuzzy membership function as well as to compare the results with standard methods. A total of five datasets related to classification problems obtained from the UCI Machine Learning Repository were used. The results showed that the proposed method had higher accuracy than standard imputation methods. Moreover, triangular method performed better than Gaussian fuzzy membership function. This showed that the combination of nearest neighbor data and fuzzy membership function was more effective in handling missing values and improving classification accuracy.

RESUMEN

Population overlap and the variation within and among populations have been globally observed but is often difficult to quantify. To achieve this, numerous different methods need to be explored and validated to assist with the creation of an accurate biological profile. The current lack of databases for postcranial macromorphoscopic traits indicates the need to further investigate if the method can be employed repeatably in a forensic context. The current study aimed to assess the prevalence of eleven postcranial macromorphoscopic traits in a South African sample. A total of 271 postcrania of adult black, coloured, and white South Africans were assessed. The intra- and inter-observer agreement ranged from fair to almost perfect except for the accessory transverse foramen of C1, which had poor agreement between observers. Only seven traits differed significantly between at least two of the groups. Univariate and multivariate random forest models were created to test the positive predictive performance of the traits to classify population affinity. The classification accuracies for the univariate models ranged from 33.3% to 53.0% and ranged from 54.6% to 62.1% for the multivariate models. Based on the variable importance, the traits assessing spinous process bifurcation were the most discriminatory variables. The results indicate that the postcranial MMS approach does not outperform current methods employed to estimate population affinity. Further research needs to be done for the method to have practical applicability for medicolegal casework in South Africa.

Asunto(s)

RESUMEN

PURPOSE: This study introduces an algorithm specifically designed for processing unprocessed 12-lead electrocardiogram (ECG) data, with the primary aim of detecting cardiac abnormalities. METHODS: The proposed model integrates Diagonal State Space Sequence (S4D) model into its architecture, leveraging its effectiveness in capturing dynamics within time-series data. The S4D model is designed with stacked S4D layers for processing raw input data and a simplified decoder using a dense layer for predicting abnormality types. Experimental optimization determines the optimal number of S4D layers, striking a balance between computational efficiency and predictive performance. This comprehensive approach ensures the model's suitability for real-time processing on hardware devices with limited capabilities, offering a streamlined yet effective solution for heart monitoring. RESULTS: Among the notable features of this algorithm is its strong resilience to noise, enabling the algorithm to achieve an average F1-score of 81.2% and an AUROC of 95.5% in generalization. The model underwent testing specifically on the lead II ECG signal, exhibiting consistent performance with an F1-score of 79.5% and an AUROC of 95.7%. CONCLUSION: It is characterized by the elimination of pre-processing features and the availability of a low-complexity architecture that makes it suitable for implementation on numerous computing devices because it is easily implementable. Consequently, this algorithm exhibits considerable potential for practical applications in analyzing real-world ECG data. This model can be placed on the cloud for diagnosis. The model was also tested on lead II of the ECG alone and has demonstrated promising results, supporting its potential for on-device application.

Asunto(s)

RESUMEN

OBJECTIVE: Major depressive disorder (MDD) and bipolar disorder (BD) are serious chronic disabling mental and emotional disorders, with symptoms that often manifest atypically in children and adolescents, making diagnosis difficult without objective physiological indicators. Therefore, we aimed to objectively identify MDD and BD in children and adolescents by exploring their voiceprint features. METHODS: This study included a total of 150 participants, with 50 MDD patients, 50 BD patients, and 50 healthy controls aged between 6 and 16 years. After collecting voiceprint data, chi-square test was used to screen and extract voiceprint features specific to emotional disorders in children and adolescents. Then, selected characteristic voiceprint features were used to establish training and testing datasets with the ratio of 7:3. The performances of various machine learning and deep learning algorithms were compared using the training dataset, and the optimal algorithm was selected to classify the testing dataset and calculate the sensitivity, specificity, accuracy, and ROC curve. RESULTS: The three groups showed differences in clustering centers for various voice features such as root mean square energy, power spectral slope, low-frequency percentile energy level, high-frequency spectral slope, spectral harmonic gain, and audio signal energy level. The model of linear SVM showed the best performance in the training dataset, achieving a total accuracy of 95.6% in classifying the three groups in the testing dataset, with sensitivity of 93.3% for MDD, 100% for BD, specificity of 93.3%, AUC of 1 for BD, and AUC of 0.967 for MDD. CONCLUSION: By exploring the characteristics of voice features in children and adolescents, machine learning can effectively differentiate between MDD and BD in a population, and voice features hold promise as an objective physiological indicator for the auxiliary diagnosis of mood disorder in clinical practice.

RESUMEN

Objective: To evaluate the latent structure, internal consistency, convergent and discriminant validity, diagnostic accuracy, and criterion validity of the Montreal Cognitive Assessment's auditory items (MoCA-22), which has previously been evaluated in small samples if at all. Methods: 11,284 participants completed the MoCA over 1-2 visits to an Alzheimer Disease Research Center (Mage = 69.2, Meducation = 15.9, 57.6% women, 92.4% non-Hispanic white). MoCA-22 items were probed with alpha, omega, confirmatory factor analysis, and test-retest correlations. Scores were related to measures of neurocognition, daily functioning, behavioral-psychological symptoms (BPS), and vision performance for convergent-discriminant and criterion validity. Dementia stage was used to calculate area under the receiver operating characteristic (AUC-ROC) curves and cutoffs for mild cognitive impairment (MCI) and dementia. Results: A single-factor had good fit (CFI = .961; TLI = .945; RMSEA = .061; SRMR = .031), with good internal consistency (Omega total = .83) and test-retest consistency (ICC = .92 at 2.7 years). The strongest convergent correlations were with general cognition and executive functioning, while discriminant validity was demonstrated with its weakest and negative correlations being with BPS. There was strong classification accuracy in distinguishing MCI from normal cognition (AUC = .79; optimal cutoff point < 18), and mild-to-moderate dementia from MCI (AUC = .85; optimal cutoff point < 13). Furthermore, the MoCA-22 had negligible-to-small differences among those with and without vision limitations. Conclusions: These findings add to the evidence of the MoCA-22's utility and it serves as a useful cognitive screening tool with sound reliability and validity.

Asunto(s)

RESUMEN

The differential of leukocytes functions as the first indicator in clinical examinations. However, microscopic examinations suffered from key limitations of low throughputs in classifying leukocytes while commercially available hematology analyzers failed to provide quantitative accuracies in leukocyte differentials. A home-developed imaging and impedance flow cytometry of microfluidics was used to capture fluorescent images and impedance variations of single cells traveling through constrictional microchannels. Convolutional and recurrent neural networks were adopted for data processing and feature extractions, which were then fused by a support vector machine to realize the four-part differential of leukocytes. The classification accuracies of the four-part leukocyte differential were quantified as 95.4% based on fluorescent images plus the convolutional neural network, 90.3% based on impedance variations plus the recurrent neural network, and 99.3% on the basis of fluorescent images, impedance variations, and deep neural networks. Based on single-cell fluorescent imaging and impedance variations coupled with deep neural networks, the four-part leukocyte differential can be realized with almost 100% accuracy.

Asunto(s)

RESUMEN

In this paper, we introduce a new and advanced multi-feature selection method for bacterial classification that uses the salp swarm algorithm (SSA). We improve the SSA's performance by using opposition-based learning (OBL) and a local search algorithm (LSA). The proposed method has three main stages, which automate the categorization of bacteria based on their unique characteristics. The method uses a multi-feature selection approach augmented by an enhanced version of the SSA. The enhancements include using OBL to increase population diversity during the search process and LSA to address local optimization problems. The improved salp swarm algorithm (ISSA) is designed to optimize multi-feature selection by increasing the number of selected features and improving classification accuracy. We compare the ISSA's performance to that of several other algorithms on ten different test datasets. The results show that the ISSA outperforms the other algorithms in terms of classification accuracy on three datasets with 19 features, achieving an accuracy of 73.75%. Additionally, the ISSA excels at determining the optimal number of features and producing a better fit value, with a classification error rate of 0.249. Therefore, the ISSA method is expected to make a significant contribution to solving feature selection problems in bacterial analysis.

RESUMEN

The measurement of psychological constructs is frequently based on self-report tests, which often have Likert-type items rated from "Strongly Disagree" to "Strongly Agree". Recently, a family of item response theory (IRT) models called IRTree models have emerged that can parse out content traits (e.g., personality traits) from noise traits (e.g., response styles). In this study, we compare the selection validity and adverse impact consequences of noise traits on selection when scores are estimated using a generalized partial credit model (GPCM) or an IRTree model. First, we present a simulation which demonstrates that when noise traits do exist, the selection decisions made based on the IRTree model estimated scores have higher accuracy rates and have less instances of adverse impact based on extreme response style group membership when compared to the GPCM. Both models performed similarly when there was no influence of noise traits on the responses. Second, we present an application using data collected from the Open-Source Psychometrics Project Fisher Temperament Inventory dataset. We found that the IRTree model had a better fit, but a high agreement rate between the model decisions resulted in virtually identical impact ratios between the models. We offer considerations for applications of the IRTree model and future directions for research.

RESUMEN

With the continuous expansion of brain-computer communication, the precise identification of brain signals has become an essential task for brain-computer equipment. However, existing classification methods are primarily concentrated on the extraction features of brain signals and obtain unacceptable performance when directly used the model to a new brain signals data, which is caused by the different people has extraordinary brain signals. In this work, we utilize the deep learning methods not only extract the features of brain signals but also learn the order information of brain signals, which can satisfy the universal brain signals. Indeed, we utilize the classification features dimension distance loss function to optimize the proposed model and enhance the classification accuracy and we compare our model with existing classification methods to evaluate proposed model. From our extensive experimental results and analysis, we can conclude that our model can achieve the classification of brain signals with the reasonable accuracy and acceptable costs.

Asunto(s)

RESUMEN

There has been a tremendous increase in the popularity of social media such as blogs, Instagram, twitter, online websites etc. The increasing utilization of these platforms have enabled the users to share information on a regular basis and also publicize social events. Nevertheless, most of the multimedia events are filled with social bots which raise concerns on the authenticity of the information shared in these events. With the increasing advancements of social bots, the complexity of detecting and fact-checking is also increasing. This is mainly due to the similarity between authorized users and social bots. Several researchers have introduced different models for detecting social bots and fact checking. However, these models suffer from various challenges. In most of the cases, these bots become indistinguishable from existing users and it is challenging to extract relevant attributes of the bots. In addition, it is also challenging to collect large scale data and label them for training the bot detection models. The performance of existing traditional classifiers used for bot detection processes is not satisfactory. This paper presents:•A machine learning based adaptive fuzzy neuro model integrated with a hist gradient boosting (HGB) classifier for identifying the persisting pattern of social bots for fake news detection.•And Harris Hawk optimization with Bi-LSTM for social bot prediction.•Results validate the efficacy of the HGB classifier which achieves a phenomenal accuracy of 95.64 % for twitter bot and 98.98 % for twitch bot dataset.

RESUMEN

Remote sensing images are important data sources for land cover mapping. As one of the most important artificial features in remote sensing images, buildings play a critical role in many applications, such as population estimation and urban planning. Classifying buildings quickly and accurately ensures the reliability of the above applications. It is known that the classification accuracy of buildings (usually indicated by a comprehensive index called F1) is greatly affected by image quality. However, how image quality affects building classification accuracy is still unclear. In this study, Boltzmann entropy (an index considering both compositional and configurational information, simply called BE) is employed to describe image quality, and the potential relationships between BE and F1 are explored based on images from two open-source building datasets (i.e., the WHU and Inria datasets) in three cities (i.e., Christchurch, Chicago and Austin). Experimental results show that (1) F1 fluctuates greatly in images where building proportions are small (especially in images with building proportions smaller than 1%) and (2) BE has a negative relationship with F1 (i.e., when BE becomes larger, F1 tends to become smaller). The negative relationships are confirmed using Spearman correlation coefficients (SCCs) and various confidence intervals via bootstrapping (i.e., a nonparametric statistical method). Such discoveries are helpful in deepening our understanding of how image quality affects building classification accuracy.

RESUMEN

(1) Background: The Wechsler intelligence scales are very popular in clinical practice and for research purposes. However, they are time consuming to administer. Therefore, researchers and psychologists have explored the possibility of shorter test battery compositions. (2) Methods: In this study, we investigated 13 potential short forms of the Indonesian version of the Wechsler Adult Intelligence Scale-Fourth Edition (WAIS-IV-ID). An existing standardization data set of 1745 Indonesian participants collected for the validation of the WAIS-IV-ID was used to examine the short forms' validity. These ranged from 2-subtest versions to 7-subtest versions. Regression analyses with goodness-of-fit measures were performed, and regression equations were determined for each short form to estimate the Full Scale IQ (FSIQ) score. Discrepancies between the FSIQ and the estimated FSIQ (FSIQEst) scores were examined and classification accuracies were calculated for each short form (% agreement of intelligence classification between the FSIQEst and FSIQ). (3) Results: None of the 13 short form FSIQEst values significantly differed from the FSIQ scores based on the full WAIS-IV-ID, and strong correlations were observed between each of these values. The classification accuracies of the short forms were between 56.8% and 81.0%. The 4-subtest short form of the WAIS-IV-ID consisting of the subtests Matrix Reasoning, Information, Arithmetic, and Coding had the optimal balance between best classification values and a short administration duration. The validity of this short form was demonstrated in a second study in an independent sample (N = 20). (4) Conclusions: Based on the results presented here, the WAIS-IV-ID short forms are able to reliably estimate the FSIQ, with a significant shorter administration duration. The WAIS-IV-ID short form consisting of four subtests, Matrix Reasoning, Information, Arithmetic, and Coding, was the best version according to our criteria.

RESUMEN

Much research has focused on neurodegeneration in aging and Alzheimer's disease (AD). We developed Scoring by Nonlocal Image Patch Estimator (SNIPE), a non-local patch-based measure of anatomical similarity and hippocampal segmentation to measure hippocampal change. While SNIPE shows enhanced predictive power over hippocampal volume, it is unknown whether SNIPE is more strongly associated with group differences between normal controls (NC), early MCI (eMCI), late (lMCI), and AD than hippocampal volume. Alzheimer's Disease Neuroimaging Initiative older adults were included in the first analyses (N = 1666, 513 NCs, 269 eMCI, 556 lMCI, and 328 AD). Sub-analyses investigated amyloid positive individuals (N = 834; 179 NC, 148 eMCI, 298 lMCI, and 209 AD) to determine accuracy in those on the AD trajectory. We compared SNIPE grading, SNIPE volume, and Freesurfer volume as features in seven different machine learning techniques classifying participants into their correct cohort using 10-fold cross-validation. The best model was then validated in the Australian Imaging, Biomarker & Lifestyle Flagship Study of Ageing (AIBL). SNIPE grading provided the highest classification accuracy for all classifications in both the full and amyloid positive sample. When classifying NC:AD, SNIPE grading provided an 89% accuracy (full sample) and 87% (amyloid positive sample). Freesurfer volume provided much lower accuracies of 65% (full sample) and 46% (amyloid positive sample). In the AIBL validation cohort, SNIPE grading provided a 90% classification accuracy for NC:AD. These findings suggest SNIPE grading provides increased classification accuracy over both SNIPE and Freesurfer volume. SNIPE grading offers promise to accurately identify people with and without AD.

Asunto(s)