Text Regions Research Articles

Influence of spherical protruded and dimple texture on the journal bearing performance: A comparative theoretical analysis implementing JFO boundary conditions

PurposeThe purpose of this paper is to compare the spherical protruded and dimple textured journal bearings performance characteristics with the untextured bearing.Design/methodology/approachThe governing Reynolds equation considering the mass conserving (JFO) boundary conditions is solved using the computationally efficient progressive mesh densification (PMD) method. The central difference scheme is used for the discretization of the governing Reynolds equation. The numerical code developed is validated with the experimental results available in the literature.FindingsFrom this numerical study, it has been observed that the protruded textured journal bearing gives better performance compared to the dimple textured and untextured journal bearing for friction variable, whereas dimple textured journal bearing provides better performance compared to the protruded textured and untextured journal bearing for load carrying capacity and flow coefficient. For better performance, dimple and protruded textured bearings must be textured in second-half textured region configuration.Practical implicationsThe results shown here would be quite useful for the researchers generally and the bearing designers particularly.Originality/valueThe load carrying capacity should be maximum whereas, the friction coefficient should be minimum and also the adequate flow of lubricant is necessary to maintain hydrodynamic lubrication and to remove the heat generated within the bearing due to friction, which impacts the bearing performance and life. Thus, this study would be significant in effective bearing design aspect.

Local Resultant Gradient Vector Difference and Inpainting for 3D Text Detection in the Wild

Three-dimensional (3D) text appearing in natural scene images is common due to 3D cameras and the capture of text from different angles, which presents new problems for text detection. This is because of the presence of depth information, shadows, and decorative characters in the images. In this work, we consider those images where 3D text appears with depth, as well as shadow information for text detection. We propose a novel method based on local resultant gradient vector difference (LRGVD), inpainting and a deep learning model for detecting 3D as well as two-dimensional (2D) texts in natural scene images. The boundary of components that are invariant to the above challenges is detected by exploring LRGVD. The LRGVD uses gradient magnitude and direction in a novel way for detecting the boundary of the components. Further, we propose an inpainting method in a new way for restoring the character background information using boundaries. For a given region and the input image, the inpainting method divides the whole image into planes and then propagates the values in the planes into the missing region based on posterior probabilities and neighboring information. This results in text regions with false positives. Then, the differential binarization network (DB-Net) is proposed for detecting text irrespective of orientation, background, 3D or 2D, etc. Experiments conducted on our 3D text images and standard datasets of natural scene text images, namely ICDAR 2019 MLT, ICDAR 2019 ArT, DAST1500, Total-Text and SCUT-CTW1500, show that the proposed method is effective in detecting 3D and 2D texts in the images.

Stereo matching algorithm based on combined matching cost computation and edge preserving filters

The stereo matching <span>process is one of the key areas that impact the stereo vision technologies which are commonly used in the application of three-dimensional reconstructions. The accuracy of the depth information used in three-dimensional reconstruction is directly proportional to the accuracy of the disparity obtained from stereo matching. The challenging issue in the stereo matching process is to determine the accurate corresponding point between the left image and right image, especially for image pairs that have different exposure such as different illumination and image pair with less texture region. In order to increase the accuracy of disparity value, a new stereo matching algorithm is proposed based on the combination of Sum of absolute different and census transform at matching cost computation. guided filter was used in the matching cost aggregation in order to remove noise and preserve the edge of the image. In the optimization step, the winner take all strategy is used to select the minimum matching cost. Finally, a median filter is applied to the initial disparity map for refinement purposes. The experimental results show that the algorithm is effective in reducing the error and improving the accuracy of the disparity map in different illumination regions, less textured regions and different environmental exposure.</span>

Integration of Optical and SAR Imagery for Dual PolSAR Features Optimization and Land Cover Mapping

Integrating optical and synthetic aperture radar (SAR) data at a feature level has proven to be a powerful method for land cover mapping. However, blurring effects and resolution loss induced by the despeckling procedure can degrade the estimation accuracy of SAR features and therefore introduce artifacts into data fusion. This article presents a method to solve this problem using SAR features estimation driven by both optical and dual polarimetric SAR images. The incentive behind this method is that compared to a low signal-to-noise ratio (SNR) SAR image, labeling sample from the optical image with higher SNR and spectral resolution can enhance sample homogeneity. The homogeneous sample following the same population can ensure SAR feature estimation accuracy and simultaneously preserve the spatial resolution. The optimum SAR features, together with optical images, are then integrated using the machine learning algorithm to obtain a classification map. We further fully evaluate the influence of SAR features estimation on fusion classification. The effectiveness of the proposed method is validated by moderate resolution Sentinel-1/2 datasets over the Saint Louis area, where abundant land cover types and rich texture present a relative challenge for classification. The quantitative analysis indicates that the proposed method can increase around 5%–10% in overall accuracy than the state-of-the-art methods. The results confirm the necessity of SAR features with high quality for land cover classification, especially over the textural region.

Depth Map Super-Resolution Based on Dual Normal-Depth Regularization and Graph Laplacian Prior

The edge information plays a key role in the restoration of a depth map. Most conventional methods assume that the color image and depth map are consistent in edge areas. However, complex texture regions in the color image do not match exactly with edges in the depth map. In this paper, firstly, we point out that in most cases the consistency between normal map and depth map is much higher than that between RGB-D pairs. Then we propose a dual normal-depth regularization term to guide the restoration of depth map, which constrains the edge consistency between normal map and depth map back and forth. Moreover, considering the bimodal characteristic of weight distribution that exists in depth discontinuous areas, a reweighted graph Laplacian regularizer is proposed to promote this bimodal characteristic. And this regularization is incorporated into a unified optimization framework to effectively protect the piece-wise smoothness(PWS) characteristics of depth map. By treating depth image as graph signal, the weight between two nodes is adapted according to its content. The proposed method is tested for both noise-free and noisy cases, and is compared against the state-of-the-art methods on both synthesis and real captured datasets. Extensive experimental results demonstrate the superior performance of our method compared with most state-of-the-art works in terms of both objective and subjective quality evaluations. Specifically, our method is more effective on edge areas and more robust to noises.

Numerical simulation and experimental investigation on tribological performance of SiC surface with squamous groove micro texture

AbstractTo improve the tribological characteristics of the face seal, the squamous groove micro textures with different area proportion and the circular dimple micro texture were designed and manufactured. The disc‐disc friction and wear experiments and numerical simulation were carried out in oil lubrication. The pressure distribution, friction coefficient, fluid temperature and wear morphology are investigated. The study show that the friction coefficients of the micro textured and untextured samples decline with the increase of load, but first increase and then decrease with the rise of rotational speed. Micro textured samples could decline the friction coefficient effectively, and the asymmetric structure in the lubricant inflow and outflow region of the squamous groove micro texture makes it have better tribological performance. In addition, the micro texture can increase heat dissipation area, reduce heat production and collect wear debris, which could modify the lubricant temperature and wear morphology effectively.

Text Region Conditional Generative Adversarial Network for Text Concealment in the Wild

Textual information appearing on the captured image may contain personal information. In various circumstances, publishing such images in the public domain may create a threat of privacy leak. To avoid these situations, we propose a text concealment method. To accomplish this task, we have used a conditional generator, which is a text region conditioned concealment network. The text regions predicted by the detector network are used as a conditioning criterion for the concealment process. The text region prediction in the form of a word-level bounding box may contain stroke pixels as well as the background pixels. However, to reduce the number of background pixels for proper conditioning of text concealment network, a character level annotation is used for the generator in place of word-level bounding box annotation. It helps to focus more on strokes as compared to the background pixels. A character-level symmetric line representation of text has been proposed to obtain finer level text region prediction as compared to the character-level bounding box. The proposed model is trainable from end-to-end. The text region conditioned generator is trained from the loss of global and local discriminators. The proposed method is validated on public scene text image datasets such as ICDAR 2015, COCO-Text and Synthesis dataset. The proposed architecture shows competitive results as compared to other state-of-the-art approaches.

MR-FPN: Multi-Level Residual Feature Pyramid Text Detection Network Based on Self-Attention Environment.

With humanity entering the age of intelligence, text detection technology has been gradually applied in the industry. However, text detection in a complex background is still a challenging problem for researchers to overcome. Most of the current algorithms are not robust enough to locate text regions, and the problem of the misdetection of adjacent text instances still exists. In order to solve the above problems, this paper proposes a multi-level residual feature pyramid network (MR-FPN) based on a self-attention environment, which can accurately separate adjacent text instances. Specifically, the framework uses ResNet50 as the backbone network, which is improved on the feature pyramid network (FPN). A self-attention module (SAM) is introduced to capture pixel-level relations, increase context connection, and obtain efficient features. At the same time, the multi-scale enhancement module (MEM) improves the expression ability of text information, extracting strong semantic information and integrating the multi-scale features generated by the feature pyramid. In addition, information regarding the upper features will cause loss when the feature pyramid is passed down step by step, and multi-level residuals can effectively solve this problem. The proposed model can effectively improve the fusion ability of the feature pyramid, provide more refined features for text detection, and improve the robustness of text detection. This model was evaluated on CTW1500, Total-Text, ICDAR2015, and MSRA-TD500 datasets of different kinds and achieved varying degrees of improvement. It is worth mentioning that the F-measure of 83.31% obtained by this paper on the Total-Text dataset exceeds that of the baseline system by 5%.

DETECTION AND RECOGNITION OF HINDI TEXT FROM NATURAL SCENES AND ITS TRANSLITERATION TO ENGLISH

India is a country with many cultures and if you travel from one place to another, you might find yourself in totally different culture. This also means the languages change from place to place in India and it gets very difficult to read signboards, shop names and even many other common things written in local languages. This can create problems for not only the travelers travelling from other countries but also the people who move withing the country from different regions. But most of the signboards, shop names or other landmarks mostly use English or Hindi in most of the regions. Here we propose a complete text detection & recognition as well as transliteration system that will help travelers read text written in Hindi on any signboards or shops and then transliterate that detected text into English. The proposed system is capable of detecting text written in Hindi language in natural environment using Progressive Scale Expansion algorithm and then transliterating the detected text into English language. Our proposed system can detect text in tough scenarios, and it can even detect curved text from natural images. Our system after detecting text region, extracts the text from the detected area using PyTesseract OCR engine and then the extracted text is further transliterated into English text with the help of seq2seq MultiRNN LSTM model which gives us accurate transliterations without losing the actual pronunciation of the original Hindi words. We use a synthetic dataset for Hindi Text images containing approx. 100000 for Text Detection and FIRE2013 dataset for transliteration. The overall system is evaluated using BLEU score.

Tikkoun Sofrim: Making Ancient Manuscripts Digitally Accessible: The Case of Midrash Tanhuma

Making ancient handwritten manuscripts accessible to the general public is challenging, for several reasons. Foremost, they are handwritten. Each and every one is unique, so there is a need for manual transcription for providing enough examples for training a machine-learning-based algorithm to automatically transcribe the handwritten text. Moreover, the quality of the text is diverse—over time the ink faded, pages were damaged, and so forth. Furthermore, the boundaries of the textual regions on a page and the lines of text are not standard. Sometimes there are corrections above the lines, the lines are curved, there are comments and annotations on the margins, and more. A possible solution for these challenges is having a “person in the loop.” However, manual correction brings with it another challenge—how to address disagreement between annotations (as usually several corrections are considered before a decision is taken about the correct transcription). Tikkoun-Sofrim is a system that integrates automatic handwritten text recognition with manual, crowdsourced error correction, introducing an automatic decision process about when to stop asking for additional transcription and selecting the best transcription, declaring it as the recommended agreed reading. The system was applied to several manuscripts of “Midrash Tanhuma,” a medieval Hebrew rabbinic homiletic text, achieving a high level of success.

Empirical evidence that bone collagen molecules denature as a result of bone fracture

Cortical bone tissue, primarily composed of collagen, hydroxyapatite, and water, is a strong and tough natural, structural biomaterial. The integrity of the collagenous phase (native triple helix vs. damaged/denatured coil) has previously been correlated via various means, including hydrothermal isometric tension testing and FTIR and Raman spectroscopy, with the capability of cortical bone to undergo stable fracture. Collagen is a relatively stable protein, requiring around 70 J/g to thermally denature its native triple helix structure, through the melting of hydrogen bonds. It is widely thought that bone collagen molecules denature (unravel) during fracture, acting as a molecular-scale mechanical toughening mechanism, but this has not been empirically demonstrated to date. A new technology, fluorescently-labeled collagen hybridizing peptides (F–CHP), enables imaging that specifically detects denatured collagen. This provides an opportunity to empirically test whether bone collagen molecules do denature during bone fracture. Here, F–CHP was used to stain fracture surfaces produced by transverse Mode-I fracture of chevron-notched bovine and human cortical bone beams. The fracture surfaces demonstrated increased staining, above the level of rigorous paired controls, and the staining directly correlated with the work-to-fracture (WFx) of bovine bone beams. This increased denaturation signal was also constrained to a rough textured region visible on the fracture surface, which is known to correspond with stable tearing. Similar staining was also detected on the fracture surfaces of human cortical bone. Increased staining was not detected on the fracture surfaces of specimens that were dehydrated prior to fracture, suggesting a role for water in the denaturation process. This study provides the first empirical evidence of bone collagen denaturation resulting from cortical bone fracture and extends our understanding of this mechanism towards the mechanical performance of cortical bone.

USB Digital Microscope Endoscope Camera – An Effective Tool for Quick Morphological Characterization of Laser‐induced Microstructures

AbstractIn this work, the authors have reported the investigation on the use of USB camera for morphological characterization to optimize the processing parameters in laser microstructuring of Si surface. For this, microstructures are generated on a monocrystalline silicon (Si) surface using nitrogen nanosecond laser. The wavelength, pulse energy, and pulse duration of the laser are of 337 nm, 170 µJ, and ∼3.5 ns, respectively. The USB Digital Microscope Endoscope (UDME) Camera is used to capture the images of microstructures generated on the Si surface. It is a compact, portable, easy to handle, and inexpensive camera calibrated by ImageJ software. It has been used for two purposes, first, for conforming whether the microstructures are formed on the surface of Si or not in given laser processing parameters and determining the dimensions of the laser textured region. The second information is used for deciding the overlapping condition. This overlapping is mostly essential for growth of microstructures in large area surface of Si through raster scanning mode. Field emission scanning electron microscopy is done to cross verify the results obtained in UDME Camera.

Self-Supervised Stereo Matching Method Based on SRWP and PCAM for Urban Satellite Images

In this paper, we propose a self-supervised stereo matching method based on superpixel random walk pre-matching (SRWP) and parallax-channel attention mechanism (PCAM). Our method is divided into two stages, training and testing. First, in the training stage, we obtain pre-matching results of stereo images based on superpixel random walk, and some matching points with high confidence are selected as labeled samples. Then, a stereo matching network is constructed to describe the matching correlation by calculating the attention scores of any two points between different images through the parallax-channel attention mechanism, superimposing the scores of each layer to calculate the disparity. The network is trained using the labeled samples and some unsupervised constraint criteria. Finally, in the testing stage, the trained network is used to obtain stereo matching relations of stereo images. The proposed method does not need manually labeled training samples and is more suitable for 3D reconstruction under mass satellite remote sensing data. Comparative experiments on multiple datasets show that our method has a stereo matching EPE of 2.44 and a 3D reconstruction RMSE of 2.36 m. Especially in the weak texture and parallax abrupt change regions, we can achieve more advanced performance than other methods.

Scientific Computational Visual Analysis of Wood Internal Defects Detection in View of Tomography Image Reconstruction Algorithm

With the development of non-destructive testing technology and computer technology, wood non-destructive testing technology will develop towards intelligence and automation. The primary condition for development is to be able to carry out the testing of various physical properties of wood without destroying the wood itself. Wood defects refer to abnormal wood structure. Its existence will affect the quality of wood, change the normal performance of wood, and reduce the utilization rate and use value of wood. The purpose of this article is to try to find an effective detection method without damaging the original structure of the wood. It can accurately and quickly determine the defect information on the wood surface. This research mainly discusses the visual analysis of wood internal defect detection based on tomographic image reconstruction algorithm. According to the analysis of the planks to be tested, this paper determines the structural characteristics, the types of defects to be tested, and the classification standards. To analyze the principle of machine vision imaging, this paper designs a hardware experiment system for wood board imaging, by observing the collected Biyun Temple building wood images and summarizing the Biyun Temple building wood defects, surface texture features, and various appearance features in the image. Based on digital image processing technology, this paper designs a complete set of real-time timber classification and detection algorithms. The algorithm realizes the extraction of Biyun Temple’s architectural wood region, texture region extraction, and the comprehensive feature vector extraction of Biyun Temple’s architectural wood. In the image preprocessing stage, the color image is converted into a grayscale image through the grayscale processing of the image. Through the equalization processing of the histogram, the defect features in the defect image are highlighted. Through image smoothing and denoising processing, the noise points that may exist in the image are removed. The signal-to-noise ratio of the Marching Cube algorithm is maintained at a relatively high level under different noise conditions, which is about 3-4db higher than the Gaussian method with OPED. This research is helpful for the continuation and inheritance of ancient architectural attainments.

Challenge of modelling GLORIA observations of upper troposphere–lowermost stratosphere trace gas and cloud distributions at high latitudes: a case study with state-of-the-art models

Abstract. Water vapour and ozone are important for the thermal and radiative balance of the upper troposphere (UT) and lowermost stratosphere (LMS). Both species are modulated by transport processes. Chemical and microphysical processes affect them differently. Thus, representing the different processes and their interactions is a challenging task for dynamical cores, chemical modules and microphysical parameterisations of state-of-the-art atmospheric model components. To test and improve the models, high-resolution measurements of the UT–LMS are required. Here, we use measurements taken in a flight of the GLORIA (Gimballed Limb Observer for Radiance Imaging of the Atmosphere) instrument on HALO (High Altitude and LOng Range Research Aircraft). The German research aircraft HALO performed a research flight on 26 February 2016 that covered deeply subsided air masses of the aged 2015/16 Arctic vortex, high-latitude LMS air masses, a highly textured region affected by troposphere-to-stratosphere exchange and high-altitude cirrus clouds. Therefore, it provides a challenging multifaceted case study for comparing GLORIA observations with state-of-the-art atmospheric model simulations in a complex UT–LMS region at a late stage of the Arctic winter 2015/16. Using GLORIA observations in this manifold scenario, we test the ability of the numerical weather prediction (NWP) model ICON (ICOsahedral Nonhydrostatic) with the extension ART (Aerosols and Reactive Trace gases) and the chemistry–climate model (CCM) EMAC (ECHAM5/MESSy Atmospheric Chemistry – fifth-generation European Centre Hamburg general circulation model/Modular Earth Submodel System) to model the UT–LMS composition of water vapour (H2O), ozone (O3), nitric acid (HNO3) and clouds. Within the scales resolved by the respective model, we find good overall agreement of both models with GLORIA. The applied high-resolution ICON-ART set-up involving an R2B7 nest (local grid refinement with a horizontal resolution of about 20 km), covering the HALO flight region, reproduces mesoscale dynamical structures well. Narrow moist filaments in the LMS observed by GLORIA at tropopause gradients in the context of a Rossby wave breaking event and in the vicinity of an occluded Icelandic low are clearly reproduced by the model. Using ICON-ART, we show that a larger filament in the west was transported horizontally into the Arctic LMS in connection with a jet stream split associated with poleward breaking of a cyclonically sheared Rossby wave. Further weaker filaments are associated with an older tropopause fold in the east. Given the lower resolution (T106) of the nudged simulation of the EMAC model, we find that this model also reproduces these features well. Overall, trace gas mixing ratios simulated by both models are in a realistic range, and major cloud systems observed by GLORIA are mostly reproduced. However, we find both models to be affected by a well-known systematic moist bias in the LMS. Further biases are diagnosed in the ICON-ART O3, EMAC H2O and EMAC HNO3 distributions. Finally, we use sensitivity simulations to investigate (i) short-term cirrus cloud impacts on the H2O distribution (ICON-ART), (ii) the overall impact of polar winter chemistry and microphysical processing on O3 and HNO3 (ICON-ART and EMAC), (iii) the impact of the model resolution on simulated parameters (EMAC), and (iv) consequences of scavenging processes by cloud particles (EMAC). We find that changing the horizontal model resolution results in notable systematic changes for all species in the LMS, while scavenging processes play a role only in the case of HNO3. We discuss the model biases and deficits found in this case study that potentially affect forecasts and projections (adversely) and provide suggestions for further model improvements.

Gaussian Pyramid Decomposition in Copy-Move Image Forgery Detection with SIFT and Zernike Moment Algorithms

One of the easiest manipulation methods is a copy-move forgery, which adds or hides objects in the images with copies of certain parts at the same pictures. The combination of SIFT and Zernike Moments is one of many methods that helping to detect textured and smooth regions. However, this combination is slowest than SIFT individually. On the other hand, Gaussian Pyramid Decomposition helps to reduce computation time. Because of this finding, we examine the impact of Gaussian Pyramid Decomposition in copy-move detection with SIFT and Zernike Moments combinations. We conducted detection test in plain copy-move, copy-move with rotation transformation, copy-move with JPEG compression, multiple copy-move, copy-move with reflection attack, and copy-move with image inpainting. We also examine the detections result with different values of gaussian pyramid limit and different area separation ratios. In detection with plain copy-move images, it generates low level of accuracy, precision and recall of 58.46%, 18.21% and 69.39%, respectively. The results are getting worse in for copy-move detection with reflection attack and copy-move with image inpainting. This weakness happened because this method has not been able to detect the position of the part of the image that is considered symmetrical and check whether the forged part uses samples from other parts of the image.

Entropy-based kernel graph cut for textural image region segmentation

Entropy-based kernel graph cut for textural image region segmentation

HessHist: A Hessian‐matrix weighted histogram for image contrast enhancement

For image contrast enhancement operation, it is a keypoint to obtain more natural enhanced results and keep more details without distortion. In this paper, a novel image Hessian-matrix weighted histogram for image contrast enhancement is proposed, which can improve the contrast of smooth regions and simultaneously restrain the contrast of texture regions. In the proposed method, the multi-scale fractional-order Hessian-matrix is firstly utilized to detect and quantify the texture information of the input image, which explores the regions that should be contrasted or should be restrained. Then, the strong texture regions are suppressed by a designed suppress function. Finally, the information on unsuppressed regions and suppressed texture regions will be count by a histogram, which is termed as Hessian-matrix weighted Histogram (HessHist) in this paper. According to HessHist, the corresponding cumulative distribution function will realize the contrast enhancement operation on the input image. For real-time application, the integral images are introduced for fast computation of the HessHist. Experimental results show that the proposed HessHist-based image enhancement algorithm preserves more details of input image without distortion, and is competitive with state-of-the-art image enhancement algorithms in both subjective visual perception and objective evaluation metrics.

Cross-Modal Object Detection Based on a Knowledge Update.

As an important field of computer vision, object detection has been studied extensively in recent years. However, existing object detection methods merely utilize the visual information of the image and fail to mine the high-level semantic information of the object, which leads to great limitations. To take full advantage of multi-source information, a knowledge update-based multimodal object recognition model is proposed in this paper. Specifically, our method initially uses Faster R-CNN to regionalize the image, then applies a transformer-based multimodal encoder to encode visual region features (region-based image features) and textual features (semantic relationships between words) corresponding to pictures. After that, a graph convolutional network (GCN) inference module is introduced to establish a relational network in which the points denote visual and textual region features, and the edges represent their relationships. In addition, based on an external knowledge base, our method further enhances the region-based relationship expression capability through a knowledge update module. In summary, the proposed algorithm not only learns the accurate relationship between objects in different regions of the image, but also benefits from the knowledge update through an external relational database. Experimental results verify the effectiveness of the proposed knowledge update module and the independent reasoning ability of our model.

Reading order detection on handwritten documents

Recent advances in Handwritten Text Recognition and Document Layout Analysis have made it possible to convert digital images of manuscripts into electronic text. However, providing this text with the correct structure and context is still an open problem that needs to be solved to actually enable extracting the relevant information conveyed by the text. The most important structure needed for a set of text elements is their reading order. Most of the studies on the reading order problem are rule-based approaches and focus on printed documents. Much less attention has been paid so far to handwritten text documents, where the problem becomes particularly important—and challenging. In this work, we propose a new approach to automatically determine the reading order of text regions and lines in handwritten text documents. The task is approached as a sorting problem where the order-relation operator is automatically learned from examples. We experimentally demonstrate the effectiveness of our method on three different datasets at different hierarchical levels.