ISI Digital Commons (Indian Statistical Institute )
Not a member yet
7571 research outputs found
Sort by
Soft set-based MSER end-to-end system for occluded scene text detection, recognition and prediction
No full textThe presence of unpredictable occlusions on natural scene text is a significant challenge, exacerbating the difficulties already posed on text detection and recognition by the variability of such images. Addressing the need for a robust, consistently performing approach that can effectively address the above challenges, this paper presents a new Soft Set-based end-to-end system for text detection, recognition and prediction in occluded natural scene images. This is the first approach to integrate text detection, recognition and prediction, unlike existing systems developed for end-to-end text spotting (text detection and recognition) only. For candidate text components detection, the proposed combination of Soft Sets with Maximally Stable Extremal Regions (SS-MSER) improves text detection and spotting in natural scene images, irrespectively of the presence of arbitrarily orientated and shaped text, complex backgrounds and occlusion. Furthermore, a Graph Recurrent Neural Network is proposed for grouping candidate text components into text lines and for fitting accurate bounding boxes to each word. Finally, a Convolutional Recurrent Neural Network (CRNN) is proposed for the recognition of text and for predicting missing characters due to occlusion. Experimental results on a new occluded scene text dataset (OSTD) and on the most relevant benchmark natural scene text datasets demonstrate that the proposed system outperforms the state-of-the-art in text detection, recognition and prediction. The code and dataset are available at https://github.com/alloydas/Softset-MSER-Based-Occluded-Scene-Text-Spotting/blob/master/Soft_set_MSER.ipyn
SwinSight: a hierarchical vision transformer using shifted windows to leverage aerial image classification
No full textIn aerial image classification, integrating advanced vision transformers with optimal preprocessing techniques is pivotal for enhancing model performance. This study presents SwinSight, a novel hierarchical vision transformer optimized for aerial image classification, which effectively addresses the computational challenges typically associated with transformers through a shifted window mechanism. The core of the research focuses on enhancing model performance by integrating a systematic preprocessing approach using Discrete Cosine Transform (DCT), Discrete Wavelet Transform (DWT), and Fast Fourier Transform (FFT). An extensive ablation study evaluates six permutations of these techniques, aiming to identify the most effective sequence for preprocessing. Results indicate that the sequence of DCT, followed by DWT, then FFT, significantly excels, achieving a high classification accuracy of 93.16% and maintaining a rapid inference time of 0.0049 seconds per frame. This sequence’s superior performance highlights the critical role of preprocessing order in optimizing feature extraction, thereby boosting the efficacy of the classification process. SwinSight’s advancements not only set a new benchmark for aerial image analysis but also offer broader implications for enhancing image processing workflows in various applications, contributing to theoretical insights and practical improvements in image-based machine learning tasks. This paper not only offers a practical solution for aerial image classification for diverse applications such as agriculture, environmental monitoring, land use applications, security, and beyond but also presents a novel SAIOD (Sikkim Aerial Images dataset for Object Detection) to the computer vision research community, fostering added advancements
Takagi-Sugeno-Kang Fuzzy Systems for High-Dimensional Multilabel Classification
No full textMultilabel classification (MLC) refers to associating each instance with multiple labels simultaneously. MLC has gained much importance due to its ability to better reflect the complexity of the real world classification problems. Fuzzy system (FS) has excellent nonlinear modeling capability and strong interpretability, which makes it a promising model for complex MLC problems. However, it is widely known that FS suffers from the \u27curse of dimensionality.\u27 Here, an adaptive membership function (MF) along with its generalized version is proposed to address high-dimensional problems. These MFs can effectively overcome \u27numeric underflow\u27 in FS while preserving interpretability as much as possible. On this basis, a novel fuzzy rule based MLC framework called multilabel high-dimensional Takagi-Sugeno-Kang fuzzy system (ML-HDTSK FS) is proposed. This model can handle data with over ten thousand dimensionality. In addition, ML-HDTSK FS uses a decomposed label correlation learning strategy to efficiently capture both high and low levels of relationship between labels, and adopts a group L21 penalty to realize the learning of label-specific features. Combining these two new multilabel learning strategies and the novel adaptive MF, ML-HDTSK FS becomes a more powerful tool for various MLC problems. The effectiveness of ML-HDTSK FS is demonstrated on seventeen benchmark multilabel datasets, and its performance is compared with eleven MLC algorithms. The experimental results confirm the validity of the proposed ML-HDTSK FS, and demonstrate the superiority of it in dealing with MLC problems, especially for high dimensional ones
THE FRIEDRICHS OPERATOR AND CIRCULAR DOMAINS
No full textThe Friedrichs operator of a domain (in Cn) is c/ose/y re/ated to its Bergman projection and encodes crucia/ information (geometric, quadrature, potentia/ theoretic etc.) about the domain. We show that the Friedrichs operator of a domain has rank one if the domain can be covered by a circu/ar domain via a proper ho/omorphic map of finite mu/tip/icity whose Jacobian is a homogeneous po/ynomia/. As an app/ication, we show that the Friedrichs operator is of rank one on the tetrab/ock, pentab/ock, and the symmetrized po/ydisc – domains of significance in the study of μ-synthesis in contro/ theory
The unit-Gompertz distribution revisited: properties and characterizations
No full textIn a recent paper, the unit-Gompertz (UG) distribution has been introduced and some of its properties have been studied. In a follow up paper, some of the subtle errors in the original paper have been corrected and some other interesting properties of this new distribution have been studied. In the present work, some more important properties are investigated. Moreover, to the best of our knowledge, no characterization results on this distribution have appeared in the literature. These are addressed in the present paper
The WHO genomics program of work for equitable implementation of human genomics for global health
No full textUtilisation of public healthcare services by an indigenous group: A mixed-method study among Santals of West Bengal, India
No full textA barrier to meeting the goal of universal health coverage in India is the inequality in utilisation of health services between indigenous and non-indigenous people. This study aimed to explore the determinants of utilisation, or non-utilisation, of public healthcare services among the Santals, an indigenous community living in West Bengal, India. The study holistically explored the utilisation of public healthcare facilities using a framework that conceptualised service coverage to be dependent on a set of determinants - viz. the nature and severity of the ailment, availability, accessibility (geographical and financial), and acceptability of the healthcare options and decision-making around these further depends on background characteristics of the individual or their family/household. This cross-sectional study adopts ethnographic approach for detailed insight into the issue and interviewed 422 adult members of Santals living in both rural (Bankura) and urban (Howrah) areas of West Bengal for demographic, socio-economic characteristics and healthcare utilisation behaviour using pre-tested data collection schedule. The findings revealed that utilisation of the public healthcare facilities was low, especially in urban areas. Residence in urban areas, being female, having higher education, engaging in salaried occupation and having availability of private allopathic and homoeopathic doctors in the locality had higher odds of not utilising public healthcare services. Issues like misbehaviour from the health personnel, unavailability of medicine, poor quality of care, and high patient load were reported as the major reasons for non-utilisation of public health services. The finding highlights the importance of improving the availability and quality of care of healthcare services for marginalised populations because these communities live in geographically isolated places and have low affordability of private healthcare. The health programme needs to address these issues to improve the utilisation and reduce the inequality in healthcare utilisation, which would be beneficial for all segments of Indian population
Variation in body size and weight status among Hindu and Muslim Indian males born in the 1890s through the 1950s
No full textHindus and Muslims represent the two largest religions in India, and also differ in nutritional status, health-related habits and standard of living associated with economic disparities. In this context, the present study considered estimated secular changes in body size, proportions, and weight status among Hindu and Muslim Indian men. The data are from anthropological surveys in the 1970s which included measurements of height, weight and sitting height of 43,950 males 18–84 years (birth years 1891–1957). Leg length was estimated; the BMI and sitting height/height ratio were calculated. Heights of men 35 + years were adjusted for estimated height loss with age. Weight status was also classified relative to WHO criteria for the BMI. Anthropometric characteristics of the two groups were compared with MANCOVA with age and geographic region as covariates. Linear regression of height on year of birth was also used to estimate secular change in each group. Heights, weights, and BMIs tended to be, on average, greater among Muslim than Hindu men at most ages, while distributions by weight status between groups were negligible. Sitting height was greater among Muslim men but estimated leg length did not differ between groups; the sitting height/height ratio thus suggested proportionally shorter legs among Muslim men. Results of the regression analyses indicated negligible differences in secular change between groups across the total span of birth years but indicated a decline in adjusted heights of men in both groups born between 1891 through 1930s and little secular change among those born in the 1930s through 1957. The variation in heights, weights and BMIs between Muslim and Hindu men at most ages suggested variation in socio-economic status and dietary habits between the groups, whereas the negligible estimated secular changes in height between groups likely reflected economic, social, and nutritional conditions during the interval of British rule and the transition to independence
Weighted Combination of Łukasiewicz implication and Fuzzy Jaccard similarity in Hybrid Ensemble Framework (WCLFJHEF) for Gene Selection
No full textA framework is developed for gene expression analysis by introducing fuzzy Jaccard similarity (FJS) and combining Łukasiewicz implication with it through weights in hybrid ensemble framework (WCLFJHEF) for gene selection in cancer. The method is called weighted combination of Łukasiewicz implication and fuzzy Jaccard similarity in hybrid ensemble framework (WCLFJHEF). While the fuzziness in Jaccard similarity is incorporated by using the existing Gödel fuzzy logic, the weights are obtained by maximizing the average F-score of selected genes in classifying the cancer patients. The patients are first divided into different clusters, based on the number of patient groups, using average linkage agglomerative clustering and a new score, called WCLFJ (weighted combination of Łukasiewicz implication and fuzzy Jaccard similarity). The genes are then selected from each cluster separately using filter based Relief-F and wrapper based SVMRFE (Support Vector Machine with Recursive Feature Elimination). A gene (feature) pool is created by considering the union of selected features for all the clusters. A set of informative genes is selected from the pool using sequential backward floating search (SBFS) algorithm. Patients are then classified using Naïve Bayes’(NB) and Support Vector Machine (SVM) separately, using the selected genes and the related F-scores are calculated. The weights in WCLFJ are then updated iteratively to maximize the average F-score obtained from the results of the classifier. The effectiveness of WCLFJHEF is demonstrated on six gene expression datasets. The average values of accuracy, F-score, recall, precision and MCC over all the datasets, are 95%, 94%, 94%, 94%, and 90%, respectively. The explainability of the selected genes is shown using SHapley Additive exPlanations (SHAP) values and this information is further used to rank them. The relevance of the selected gene set are biologically validated using the KEGG Pathway, Gene Ontology (GO), and existing literatures. It is seen that the genes that are selected by WCLFJHEF are candidates for genomic alterations in the various cancer types. The source code of WCLFJHEF is available at http://www.isical.ac.in/~shubhra/WCLFJHEF.html
Weighted cumulative residual Kullback–Leibler divergence: properties and applications
No full textWeighted cumulative residual Kullback–Leibler information measure and its dynamic version are proposed and their properties are investigated. Also weighted cumulative Kullback–Leibler information measure along with its dynamic version is introduced. Monotonicity properties of the dynamic information measures are studied. A goodness-of-fit test is developed for exponential distribution using weighted cumulative residual Kullback–Leibler information measure based on complete and censored data. It is observed that proposed tests perform well for monotone decreasing hazard alternatives under censored data. Four data sets are analyzed for illustration