Search

Abstract Background Gestational diabetes mellitus (GDM) is a common pregnancy complication with long-term health consequences for mothers and their children. The escalating trends of GDM coupled with the growing prevalence of maternal obesity, a significant GDM risk factor projected to approach nearly 60% by 2030 in Kansas, has emerged as a pressing public health issue. Methods The aim of this study was to compare GDM and maternal obesity trends in rural and urban areas and investigate maternal demographic characteristics influencing the risk of GDM development over a 15-year period. Trend analyses and a binary logistic regression were employed utilizing 2005 to 2019 de-identified birth record vital statistics from the Kansas Department of Health and Environment (N = 589,605). Results Over the cumulative 15-year period, a higher prevalence of GDM was observed across age, race/ethnicity, education, and insurance source. Throughout this period, there was an increasing trend in both GDM and obese pre-pregnancy BMI age-adjusted prevalence, with noticeable rural-urban disparities. From 2005 to 2019, women, including Asians (OR: 2.73, 95% CI 2.58%-2.88%), American Indian or Alaskan Natives (OR: 1.58, 95%, CI 1.44-1.73%), Hispanics (OR: 1.42, 95% CI 1.37%-1.48%), women residing in rural areas (OR: 1.09, 95%, CI 1.06-1.12%), with advanced maternal age (35–39 years, OR: 4.83 95% CI 4.47%-5.22%; ≥40 years, OR: 6.36 95%, CI 5.80-6.98%), with lower educational status (less than high school, OR: 1.15, 95% CI 1.10%-1.20%; high school graduate, OR: 1.10, 95% CI 1.06%-1.13%), Medicaid users (OR: 1.10, 95% CI 1.06%-1.13%), or with an overweight (OR: 1.78, 95% CI 1.72%-1.84%) or obese (OR: 3.61, 95% CI 3.50%-3.72%) pre-pregnancy BMI were found to be at an increased risk of developing GDM. Conclusions There are persistent rural-urban and racial/ethnic disparities present from 2005 to 2019 among pregnant women in Kansas with or at-risk of GDM. There are several socioeconomic factors that contribute to these health disparities affecting GDM development. These findings, alongside with prominent rising maternal obesity trends, highlight the need to expand GDM services in a predominantly rural state, and implement culturally-responsive interventions for at-risk women.

Facial attribute classification algorithms frequently manifest demographic biases by obtaining differential performance across gender and racial groups. Existing bias mitigation techniques are mostly in-processing techniques, i.e., implemented during the classifier’s training stage, that often lack generalizability, require demographically annotated training sets, and exhibit a trade-off between fairness and classification accuracy. In this paper, we propose a technique to mitigate bias at the test time i.e., during the deployment stage, by harnessing prediction uncertainty and human–machine partnership. To this front, we propose to utilize those lowest percentages of test data samples identified as outliers with high prediction uncertainty. These identified uncertain samples at test-time are labeled by human analysts for decision rendering and for subsequently re-training the deep neural network in a continual learning framework. With minimal human involvement and through iterative refinement of the network with human guidance at test-time, we seek to enhance the accuracy as well as the fairness of the already deployed facial attribute classification algorithms. Extensive experiments are conducted on gender and smile attribute classification tasks using four publicly available datasets and with gender and race as the protected attributes. The obtained outcomes consistently demonstrate improved accuracy by up to 2% and 5% for the gender and smile attribute classification tasks, respectively, using our proposed approaches. Further, the demographic bias was significantly reduced, outperforming the State-of-the-Art (SOTA) bias mitigation and baseline techniques by up to 55% for both classification tasks. The demo shall be released on https://github.com/hashtaglensman/HumanintheLoop.

With the unprecedented mobile technology revolution, mobile devices have transcended from being the primary means of communication to an all-in-one platform. Consequently, an increasing number of individuals are accessing online services for e-commerce and banking via smartphones instead of traditional desktop computers. However, smartphones can be easily misplaced, lost, or stolen more often than other computing devices, thereby demanding effective user authentication mechanisms for device unlocking and secured transactions. Ocular biometrics has obtained significant attention from academia and industry because of its accuracy, security, and ease of use in mobile devices. Several studies have demonstrated the efficacy of deep learning models for ocular-based user authentication on smartphones. However, these high-performing models require enormous space and computational complexity due to the millions of parameters and computations involved. These requirements make their deployment on resource-constrained smartphones challenging. To this end, a handful of studies have been proposed for compact-size ocular-based deep-learning models to facilitate on-device deployment. In this paper, we conduct a thorough analysis of the existing neural network compression techniques applied as a standalone and in combination for ocular-based user authentication. Extensive experimental validation is performed on the two latest large-scale ocular biometric datasets collected using smartphones, namely, UFPR and VISOB 2.0 datasets. This study benchmarks the results of advanced compression techniques for further research and development in lightweight models for ocular-based user authentication on smartphones.

Abstract Considering challenges such as occlusions due to face coverings, ocular biometrics have risen as a potential remedy. Visible light ocular modalities use features extracted from in and around the eyes from ‘selfie’ captures for personal identification. A less‐visited ocular neighbourhood is the eyebrows region. Previous works on eyebrows have been mostly using subject‐specific verification protocols where identities overlap between the training and test sets and do not directly address the issue of doppelgangers and twins. Here, the evaluation of five deep learning models, lightCNN, ResNet, DenseNet, MobileNetV2, and SqueezeNet, for eyebrow‐based user authentication in a subject independent environment across different data sets, lighting conditions, resolutions, and facial expressions is done. Also a challenging simulated identical twins scenario in our training and testing data sets is presented. Our results show a 4.2% equal error rate (EER) and 0.993 area under the curve (AUC) on a subset of the FACES data set (154 subjects) and 6.8% EER, 0.978 AUC on a subset of the VISOB data set (350 subjects). Despite achieving promising accuracy during short‐term evaluations, the proposed methods showed lower accuracy when matching against long‐term data or when facing data captured under varying illumination and facial expressions.

Childhood and adolescent obesity is a serious health problem that is on the rise at the global level. Earlier, certain diseases such as Type 2 diabetes, high blood pressure, and heart disease affected only adults, but now they are being detected in young children as well. Several studies based on machine learning have been proposed to develop obesity prediction models or to determine key determinants of obesity for designing intervention tools. Despite having a rich and diverse set of literature on obesity prediction models, obesity rates are at an all-time high for both children and adolescents. There is a need of proper understanding and critical analysis of existing machine learning models in order to design effective strategies for curbing obesity at childhood and adolescent level. This paper surveys the growing body of recent literature on machine and deep learning models for obesity prediction by providing a coherent view (critical analysis) of the limitations of the existing systems. The taxonomy of the existing literature on obesity prediction into methods used, predicted outcome, factors used, type of datasets, and the associated purpose, is discussed for analysis of the state-of-the-art. This analysis revealed that a) prediction-focused models do not use variables from as many domains as predictor-focused models do, b) very few studies proposed gender-specific and race-specific obesity prediction models, c) lack of large-scale multimodal datasets and d) existing predictor-focused models obtain an accuracy range of [53.7%, 96%] with an optimum set of predictors. Further, computer vision-based methods for obesity prediction and interpretable techniques for understanding the outcome of the models are discussed as well. In addition, we have also identified novel research directions. The overall aim is to advance the state-of-the-art and improve the quality of discourse in this field.

Incremental learning enables the learner to accommodate new knowledge without retraining the existing model. It is a challenging task that requires learning from new data and preserving the knowledge extracted from the previously accessed data. This challenge is known as the stability-plasticity dilemma. We propose A-iLearn, a generic model for incremental learning which overcomes the stability-plasticity dilemma by carefully integrating the ensemble of base classifiers trained on new data with the current ensemble without retraining the model from scratch using entire data. We demonstrate the efficacy of the proposed A-iLearn model on spoof fingerprint detection application. One of the significant challenges associated with spoof fingerprint detection is the performance drop on spoofs generated using new fabrication materials. A-iLearn is an adaptive incremental learning model that adapts to the features of the “live” and “spoof” fingerprint images and efficiently recognizes the new spoof fingerprints and the known spoof fingerprints when the new data is available. To the best of our knowledge, A-iLearn is the first attempt in incremental learning algorithms that adapts to the properties of data for generating a diverse ensemble of base classifiers. From the experiments conducted on standard high-dimensional datasets LivDet 2011, LivDet 2013 and LivDet 2015, we show that the performance gain on new fake materials is significantly high. On average, we achieve 49.57% improvement in accuracy between the consecutive learning phases.

Quality of experience (QoE) assessment occupies a key role in various multimedia networks and applications. Recently, large efforts have been devoted to devise objective QoE metrics that correlate with perceived subjective measurements. Despite recent progress, limited success has been attained. In this paper, we provide some insights on why QoE assessment is so difficult by presenting few major issues as well as a general summary of quality/QoE formation and conception including human auditory and vision systems. Also, potential future research directions are described to discern the path forward. This is an academic and perspective article, which is hoped to complement existing studies and prompt interdisciplinary research.