ARTICLE | doi:10.20944/preprints202108.0360.v1
Subject: Mathematics & Computer Science, Artificial Intelligence & Robotics Keywords: Table detection, table localization, deep learning, Hybrid Task Cascade, Object detection, deformable convolution, deep neural networks, computer vision, scanned document images, document image analysis.
Online: 17 August 2021 (10:26:42 CEST)
Tables in the document image are one of the most important entities since they contain crucial information. Therefore, accurate table detection can significantly improve information extraction from tables. In this work, we present a novel end-to-end trainable pipeline, HybridTabNet, for table detection in scanned document images. Our two-stage table detector uses the ResNeXt-101 backbone for feature extraction and Hybrid Task Cascade (HTC) to localize the tables in scanned document images. Moreover, we replace conventional convolutions with deformable convolutions in the backbone network. This enables our network to detect tables of arbitrary layouts precisely. We evaluate our approach comprehensively on ICDAR-13, ICDAR-17 POD, ICDAR-19, TableBank, Marmot, and UNLV. Apart from the ICDAR-17 POD dataset, our proposed HybridTabNet outperforms earlier state-of-the-art results without depending on pre and post-processing steps. Furthermore, to investigate how the proposed method generalizes unseen data, we conduct an exhaustive leave-one-out-evaluation. In comparison to prior state-of-the-art results, our method reduces the relative error by 27.57% on ICDAR-2019-TrackA-Modern, 42.64% on TableBank (Latex), 41.33% on TableBank (Word), 55.73% on TableBank (Latex + Word), 10% on Marmot, and 9.67% on UNLV dataset. The achieved results reflect the superior performance of the proposed method.
ARTICLE | doi:10.20944/preprints202208.0287.v1
Subject: Mathematics & Computer Science, Artificial Intelligence & Robotics Keywords: table detection; document layout analysis; continual learning; incremental learning; experience replay
Online: 16 August 2022 (10:56:59 CEST)
The growing amount of data demands methods that can gradually learn from new samples. However, it is not trivial to continually train a network. Retraining a network with new data usually results in a known phenomenon, called “catastrophic forgetting.” In a nutshell, the performance of the model drops on the previous data by learning from the new instances. This paper explores this issue in the table detection problem. While there are multiple datasets and sophisticated methods for table detection, the utilization of continual learning techniques in this domain was not studied. We employed an effective technique called experience replay and performed extensive experiments on several datasets to investigate the effects of catastrophic forgetting. Results show that our proposed approach mitigates the performance drop by 15 percent. To the best of our knowledge, this is the first time that continual learning techniques are adopted for table detection, and we hope this stands as a baseline for future research.
ARTICLE | doi:10.20944/preprints202109.0059.v1
Subject: Mathematics & Computer Science, Artificial Intelligence & Robotics Keywords: table detection; table recognition; cascade Mask R-CNN; atrous convolution; recursive feature pyramid networks; document image analysis; deep neural networks; computer vision, object detection.
Online: 3 September 2021 (11:05:10 CEST)
Table detection is a preliminary step in extracting reliable information from tables in scanned document images. We present CasTabDetectoRS, a novel end-to-end trainable table detection framework that operates on Cascade Mask R-CNN, including Recursive Feature Pyramid network and Switchable Atrous Convolution in the existing backbone architecture. By utilizing a comparatively lightweight backbone of ResNet-50, this paper demonstrates that superior results are attainable without relying on pre and post-processing methods, heavier backbone networks (ResNet-101, ResNeXt-152), and memory-intensive deformable convolutions. We evaluate the proposed approach on five different publicly available table detection datasets. Our CasTabDetectoRS outperforms the previous state-of-the-art results on four datasets (ICDAR-19, TableBank, UNLV, and Marmot) and accomplishes comparable results on ICDAR-17 POD. Upon comparing with previous state-of-the-art results, we obtain a significant relative error reduction of 56.36%, 20%, 4.5%, and 3.5% on the datasets of ICDAR-19, TableBank, UNLV, and Marmot, respectively. Furthermore, this paper sets a new benchmark by performing exhaustive cross-datasets evaluations to exhibit the generalization capabilities of the proposed method.
ARTICLE | doi:10.20944/preprints201808.0265.v1
Subject: Mathematics & Computer Science, Geometry & Topology Keywords: writhe; chirality; nomenclature; link symmetries; link table; knot table; lattice polygons; DNA topology
Online: 15 August 2018 (05:18:48 CEST)
Proper identification of oriented knots and 2-component links requires a precise link nomenclature. Motivated by questions arising in DNA topology, this study aims to produce a nomenclature unambiguous with respect to link symmetries. For knots, this involves distinguishing a knot type from its mirror image. In the case of 2-component links, there are up to sixteen possible symmetry types for each topology. The study revisits the methods previously used to disambiguate chiral knots and extends them to oriented 2-component links with up to nine crossings. Monte Carlo simulations are used to report on writhe, a geometric indicator of chirality. There are ninety-two prime 2-component links with up to nine crossings. Guided by geometrical data, linking number and the symmetry groups of 2-component links, a canonical link diagram for each link type is proposed. All diagrams but six were unambiguously chosen ( , , , , , and ). We include complete tables for prime knots with up to ten crossings and prime links with up to nine crossings. We also prove a result on the behavior of the writhe under local lattice moves.
ARTICLE | doi:10.20944/preprints201903.0040.v1
Online: 4 March 2019 (10:33:58 CET)
This study examined the determinants of domestic savings mobilization among the rural poor in Uasin-Gishu county, Kenya. The general notion is that the subsistence farmers are too poor to save. This seems to be unfounded given the fact that they are general excluded from formal financial services and studies on poverty in the country show that the average propensity of the rural households to save is higher than the national average. What are the factors which motivate small scale farmers to save? The study was conducted on 446 table banking groups under the aegis of JOYWO, a table banking grouping in Kenya. Data was collected using structured questionnaires from members of groups under the umbrella of JOYWO. The findings of the study indicates that household income had a positive and significant effect on savings mobilization while dependency ratio had a negative and significant effect on savings mobilization. Household size was not significant. The results point to the need to expose the rural poor to informal savings and financing models expected to enhance income generating capabilities of the rural poor and lower the level dependency through government welfare funding for senior citizens and essential services for the young.
ARTICLE | doi:10.20944/preprints202202.0157.v1
Subject: Mathematics & Computer Science, Numerical Analysis & Optimization Keywords: mathematical models; Table Curve 3D; correlation coefficient
Online: 11 February 2022 (08:35:16 CET)
This article describes the methodology used to identify the mathematical model that describes the correlations between the input parameters of an experiment and the parameters being followed. As a technological process, the aerodynamic separation was chosen, respectively, the behavior of a solid particle within an ascending vertical airflow. The experimental data obtained were used to identify two parameters, the average linear velocity, and the angular velocity, and through the Table Curve 3D program was developed a mathematical model which describes the dependence between the input parameters (the shape and size of the solid particle and the velocity of the airflow) and the monitored parameters. In order to determine a single mathematical equation that describes as accurately as possible the correlation between the input variables and those obtained, a pyramid-type analysis was designed. The determination of the mathematical equation started from the number of equations generated by the Table Curve 3D program, then the equations with a correlation coefficient greater than 0.85 were chosen, and finally, the common equations were identified. Respecting the working methodology was identified one equation which has for the average linear velocity a correlation coefficient r2 between 0.88-0.99 and 0.86-0.99 for the angular velocity.
ARTICLE | doi:10.20944/preprints201710.0192.v1
Subject: Mathematics & Computer Science, General & Theoretical Computer Science Keywords: set; hash table; hash function; commutative semigroup
Online: 31 October 2017 (04:35:55 CET)
Hash tables are widely used. They rely on good quality hash functions. Popular data structure libraries either provide no hash functions or weak hash functions for sets or maps, making it impossible or impractical to use them as keys in other tables. This article presents three algorithms for hashing a set, two of which are simple to implement, practically fast, and can be combined. The quality evaluations follow the method of [1, chapter 2]. The insight that we are looking for commutative semigroups suggests that even better methods than symmetric polynomials may be found.
ARTICLE | doi:10.20944/preprints202007.0042.v1
Subject: Biology, Agricultural Sciences & Agronomy Keywords: fitness; life table; cotton bollworm; corn earworm; toxin; resistance management
Online: 5 July 2020 (04:42:07 CEST)
Insecticidal toxins from Bacillus thuringiensis (Bt) are valuable tools for pest management worldwide, contributing to the management of human disease insect vectors and phytophagous insect pests of agriculture and forestry. Here, we report the effects of dual and triple Bt toxins expressed in transgenic cotton cultivars on the fitness and demographic performance of Helicoverpa zea (Boddie), a noctuid pest known as cotton bollworm and corn earworm. Life-history traits were determined for individuals of three field populations from a region where H. zea overwintering is likely. Triple-gene Bt cotton cultivars expressing Cry and Vip3Aa toxins killed 100% of the larvae in all populations tested. In contrast, dual-gene Bt cotton expressing Cry1Ac+Cry1F and Cry1Ac+Cry2Ab2 allowed population growth with the intrinsic rate of population growth (rm) 38% lower than on non-Bt cotton. The insects feeding on Bt cotton plants expressing Cry1Ac+Cry2Ab2, Cry1Ac+Cry1F, or Cry1Ab+Cry2Ae exhibited reduced larval weight, survival rate, and increased development time. Additionally, fitness parameters varied significantly among the insect populations, even on non-Bt cotton plants, likely because of their different genetic background and/or previous Bt toxin exposure. This is the first report of the comparative fitness of H. zea field populations on dual-gene Bt cotton after the recent reports of field resistance to certain Bt toxins. These results document the population growth rates of H. zea from an agricultural landscape with 100% Bt cotton cultivars. Our results will help to refine models designed to predict resistance evolution and improve insect resistance management for Bt crops.
ARTICLE | doi:10.20944/preprints202103.0520.v1
Subject: Engineering, Automotive Engineering Keywords: Intuitive learning; Dynamic response; Small-scale model; Image-recognition; Shaking table
Online: 22 March 2021 (11:21:47 CET)
In the last years, more and more studies highlight the advantages of complementing traditional master classes with additional activities that improve students´ learning experience. This combination of teaching techniques is specially advised in the field of structural engineering, where intuition of the structural response it is of vital importance to understand the studied concepts. This paper deals with the introduction of a new (and more encouraging) educational tool to introduce intuitively students in the dynamic response of structures excited with an educational shaking table. Most of the educational structural health monitoring systems use sensors to determine the dynamic response of the structure. The proposed tool is based on a radically different approach, as it is based on low-cost image-recognition techniques. In fact, it only requires the use an amateur camera, a black background and a computer. In this study, the effects of both the camera location and the image quality are also evaluated. Finally, to validate the applicability of the proposed methodology, the dynamic response of small-scale buildings with different typologies is analyzed. In addition, a series of surveys were conducted in order to evaluate the activity based on student´s satisfaction and the actual acquisition and strengthening of knowledge.
ARTICLE | doi:10.20944/preprints202103.0145.v1
Subject: Engineering, Other Keywords: Stone pagoda; Masonry structure; Shaking table test; Earthquake resistance; Seismic behavior
Online: 4 March 2021 (09:15:13 CET)
In general, the stone pagoda structures with discontinuous surfaces are vulnerable to lateral forces and are severely damaged by earthquakes. After the Gyeongju earthquake in 2016 and the Pohang earthquake in 2017, the earthquakes damaged numerous stone pagoda structures due to slippage, rotation and the separation of stacked stone. To evaluate seismic resistance of masonry stone pagoda structure, we analyzed the seismic behavior of stone pagoda structure using shaking table test. Shaking frequency, permanent displacement, maximum acceleration, rocking, and sliding were assessed. Responses to simulations of the Bingol, Gyeongju, and Pohang earthquakes based on Korean seismic design standard (KDS 41 17 00) were analyzed for return periods of 1,000 and 2,400 years. We found that the type of stylobate affected the seismic resistance of stone pagoda structure. When the stylobates were stiff, seismic energy was transferred from lower to upper regions of the stone pagoda, which mainly resulted in deformation of the upper region. When the stylobates were weak, earthquake energy was absorbed in the lower regions; this was associated with large stylobate deformations. The lower part of tower body was mainly affected by rocking, because the structural members were slender. The higher part of the stone pagoda was mainly affected by sliding, because the load and contact area decreased with height.
SHORT NOTE | doi:10.20944/preprints202011.0267.v1
Subject: Earth Sciences, Atmospheric Science Keywords: Siberian fir; regression model; forest type group; bonitet; growth rate table
Online: 9 November 2020 (09:11:44 CET)
The paper presents an assessment of the growth dynamics of the modal fir plantations in the Lower Angara region. At present, a vast area of fir forests in the Lower Angara region is characterised by a significant decrease in sustainability due to periodic forest fires, insect pests outbreaks and diseases, which lead to their natural degradation and death. However, the intensity of coniferous stand growth in certain forest site characteristics persists in the long term. Therefore, creating regression models of forest growth and development involving the identification of site conditions is very important both from a practical point of view and for environmental monitoring. The materials of the mass inventory of 3491 stands served as the initial data for studying the processes of fir plantations natural growth. The Hoerl Model function is suitable for the best approximation of stand growth since it is characterised by high levelling factor (from 0.970 to 0.987) and a small standard error (not exceeding 7%). As a result of the research, there have been constructed sketches of the growth rate tables for the modal Siberian fir stands of the third bonitet class of the forb and mossy groups of forest types.
ARTICLE | doi:10.20944/preprints202109.0021.v1
Subject: Mathematics & Computer Science, Probability And Statistics Keywords: Effect size; correlation coefficient; association measure; covariance; mean square contingency coefficient; mean square effect half-size; Pearson’s Phi; 2 × 2 table; binary crosstab; gross crosstab; contingency table
Online: 1 September 2021 (14:28:47 CEST)
Evidence-based medicine (EBM) is in crisis, in part due to bad methods, which are understood as misuse of statistics that is considered correct in itself. This article exposes two related common misconceptions in statistics, the effect size (ES) based on correlation (CBES) and a misconception of contingency tables (MCT). CBES is a fallacy based on misunderstanding of correlation and ES and confusion with 2 × 2 tables, which makes no distinction between gross crosstabs (GCTs) and contingency tables (CTs). This leads to misapplication of Pearson’s Phi, designed for CTs, to GCTs and confusion of the resulting gross Pearson Phi, or mean-square effect half-size, with the implied Pearson mean square contingency coefficient. Generalizing this binary fallacy to continuous data and the correlation in general (Pearson’s r) resulted in flawed equations directly expressing ES in terms of the correlation coefficient, which is impossible without including covariance, so these equations and the whole CBES concept are fundamentally wrong. MCT is a series of related misconceptions due to confusion with 2 × 2 tables and misapplication of related statistics. The misconceptions are threatening because most of the findings from contingency tables, including CBES-based meta-analyses, can be misleading. Problems arising from these fallacies are discussed and the necessary changes to the corpus of statistics are proposed resolving the problem of correlation and ES in paired binary data. Since exposing these fallacies casts doubt on the reliability of the statistical foundations of EBM in general, we urgently need to revise them.
Subject: Physical Sciences, Acoustics Keywords: Hyperspectral Imaging, Phenolics, Anthocyanin, Table Grapes, Total Soluble Solid, PLS, MLR, Model.
Online: 1 February 2021 (12:35:56 CET)
Table grape quality is of importance for consumers and thus for producers. The objective quality determination is usually destructive and very simple with the assessment of only a couple of parameters. This study proposed to evaluate the possibility of hyperspectral imaging to characterize table grapes quality through its sugar, total flavonoid and total anthocyanin contents. Different pre-treatments (WB, SNV, 1st and 2nd derivative) and different methods were tested: PLS with full spectra, then Multiple Linear Regression (MLR) were realized after selecting the optimal wavelengths thanks to the regression coefficients (-coefficients) and the Variable Importance in Projection (VIP) scores from the full spectra. All models were good showing that hyperspectral imaging is a relevant method to assess sugar content and global phenolic content. The best model was dependent on the variable. The best models were from the full spectra and with the 2nd derivative pre-treatment for TSS; from VIPs optimal wavelengths using SNV pre-treatment for Total Flavonoid and total Anthocyanin content. Thus, relevant models were proposed using the full spectra, as well as specific windows and wavelengths in order to reduce the data sets and limit the data storage to enable an industrial use.
ARTICLE | doi:10.20944/preprints202110.0226.v1
Subject: Life Sciences, Biotechnology Keywords: kefir grains; red table grapes; kefir-like beverage; fed-batch fermentation; volatile compounds
Online: 15 October 2021 (14:04:51 CEST)
The aim of this work was to study the production of kefir-like beverage by fed-batch fermentation of red table grape juice at initial pHs of 3.99 (fermentation A) and 5.99 (fermentation B) with kefir grains during four repeated 24-h fed-batch subcultures. However, all kefir-like beverages (KLB) were characterized by low alcoholic grade (≤ 3.6%, v/v) and lactic and acetic acid concentrations. The beverages obtained from fermentation B had lower concentrations of sugars and higher microbial counts than the KLB obtained in fermentation A. In addition, the KLB from fermentation B were the most aromatic and had the highest contents in alcohols, esters, aldehydes and organic acids compared to the non-fermented juice and KLB from fermentation A. These results indicate the possibility of obtaining red table grapes KLB with their own distinctive aromatic characteristics and a high content in probiotic viable cells, contributing to the valorization of this fruit.
ARTICLE | doi:10.20944/preprints201912.0008.v1
Subject: Biology, Ecology Keywords: microclimate; water table depth; climate change impacts; cape restionaceae; species distribution modelling; maxent
Online: 2 December 2019 (10:14:59 CET)
The Cape Restionaceae species, an endemic of the Fynbos Biome, is threatened by urbanization, alien plant invasion, agricultural expansion, and groundwater extraction. This is further worsened by the semi-arid conditions and hydrological variability factors, which influences species niche dynamics. Therefore, it is important to assess and monitor the Restionaceae species for preservation of their endemism and richness. This study models the hydrological niche and distribution changes of Restionaceae species at the New Years Peak (NYP) at microclimate level for biodiversity conservation. MaxEnt modelling and GIS analytical approaches were applied at various stages in niche modelling process as follows: (i) microclimatic input raster layers’ generation, (ii) ecological modelling and hydrological niche manipulation, and (iii) spatial distributional change mapping. The hydrological niches of the Restionaceae were effectively examined under the recent climate and compared with RCP2.6 and RCP8.5 future climate scenarios as the microscale environmental inputs. The results showed that most of the studied Restionaceae species positioned themselves along a hydrological gradient. Each species tolerated a range of hydrological conditions, which formed their hydrological niche. Changing climate would cause both positive and negative species range shifts. The study assists in plant species conservation and future climate change impact analysis on endangered plant species.
ARTICLE | doi:10.20944/preprints201807.0373.v1
Subject: Medicine & Pharmacology, Sport Sciences & Therapy Keywords: table tennis; hue-saturation-value (HSV) image segmentation; low light source; automatic tracking
Online: 20 July 2018 (04:36:28 CEST)
In table tennis competitions the rule violation judgment with the greatest controversy is the height of the ball serve. This is because inaccuracy in the ball height judgment, which results in erroneous judgment, is unavoidable. Thus, we designed an automatic image judgment auxiliary system for table tennis ball height during service in this study. We used a high-speed camera to record the ball toss in the table tennis service. The designed algorithm architecture can automatically search for the ball and the position of the hand action under low light source conditions. The algorithm is mainly divided into hue-saturation-value color space processing and morphology processing using Hough transform to search for the circular ball. Experiment result shows that color segmentation can successfully and accurately determine the ball position under low light conditions. The morphology method can find the position of the hand and help determine the moment when the ball leaves the hand during the service ball toss. Finally, the actual size of the target is used to estimate the actual distance unit represented by the image pixel.
REVIEW | doi:10.20944/preprints202204.0047.v1
Subject: Mathematics & Computer Science, Computational Mathematics Keywords: Artificial Intelligence; Bottom-up Parser; Context-Free-Grammar; English Grammar; Python; Parse Table; Semantic Parser; Top-downParser
Online: 6 April 2022 (12:42:37 CEST)
The objective of parsing is to transform a natural language sentence it in to a standard order. and in a same way a sentence is tokenized with an appropriate format. There are certain English grammar evaluation rules and the parsing approach which is to be followed for the proper formation of a particular sentence syntactically and semantically using the parsing approach. A sentence in English language is the main element in the semantic parser, which creates a parse tree with the help of applying semantic dating technique to a number of phrases. A parser divides a token into smaller components by applying sets of guidelines that characterize and a series of the tokens to determine its structure of the language, which specified by grammar. The illustration provides easy records on grammatical connections, which can simply know and put into practice with those who have no prior knowledge of the language, such as those who need to obtain textual family members. The semantic family members represent the relationships of a number of the words in the sentence. We advocate utilizing our parser to acquire the tagged sets as well as a context-free layout grammatical representation for the source form. All pronouns, adverbs, singular, plural, nouns, verbs, people, adjectives, tenses and other words are kept in a database.
ARTICLE | doi:10.20944/preprints201806.0464.v1
Subject: Engineering, Mechanical Engineering Keywords: harmonic identification; adaptive linear neutral network; least mean M-estimate; electro-hydraulic servo shaking table; harmonic distortion
Online: 28 June 2018 (10:55:10 CEST)
Since the electro-hydraulic servo shaking table exists many nonlinear elements, such as, dead zone, friction and blacklash, its acceleration response has higher harmonics which result in acceleration harmonic distortion, when the electro-hydraulic system is excited by sinusoidal signal. For suppressing the harmonic distortion and precisely identify harmonics, a combination of the adaptive linear neural network and least mean M-estimate (ADALINE-LMM), is proposed to identify the amplitude and phase of each harmonic component. Namely, the Hampel’s three-part M-estimator is applied to provide thresholds for detecting and suppressing the error signal. Harmonic generators are used by this harmonic identification scheme to create input vectors and the value of the identified acceleration signal is subtracted from the true value of the system acceleration response to construct the criterion function. The weight vector of the ADALINE is updated iteratively by the LMM algorithm, and the amplitude and phase of each harmonic, even the results of harmonic components, can be computed directly online. The simulation and experiment are performed to validate the performance of the proposed algorithm. According to the experiment result, the above method of harmonic identification possesses great real-time performance and it has not only good convergence performance but also high identification precision.