Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Reconstruction of an underwater object from a sequence of images distorted by moving water waves is a challenging task. Most of the environmental research has been employing image data in recent days. The precision of this research is often dependent on the superiority of image data. In the existing approaches, the problem of analyzing video sequences when the water surface is disturbed by waves. The water waves will affect the appearance of the individual video frames such that no single frame is completely free of geometric distortion. Thus, the image acquisition from the environmental condition is more complex and crucial, but it must be focused on getting the high spectral and spatial quality. The primary intent of this paper is to plan for the intelligent higher-order spectral analysis for recovering the images from the moving water surface. The three main phases of the proposed image recovery model are (a) image pre-processing, (b) lucky region selection, and (c) image recovery. Once the pre-processing of the image is carried out, the lucky region selection is performed by computing the dice coefficient method. As a modification to the existing methods, the proposed model adopts optimized bispectra to enhance the quality of the recovered image. A hybrid algorithm with Dragonfly-Colliding Body Optimization (D-CBO) is used for enhancing the bispectra method. The proposed model has been tested on distorted underwater images. From the experimental analysis, in terms of PSNR measure, the suggested D-CBO-bispectra gets better efficiency than other conventional models, in which D-CBO-bispectra is 10.7%, 8.7%, 19%, 6.8% and 5% progressed than Blind deconv, Bispectra, Bispectra with Fourier, and Radon transform, respectively. Finally, the comparison of the proposed model with the existing approaches proves the method's efficiency. PubDate: 2022-05-17
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: The Electric Resistance Tomograph (ERT) is a customized tree specific novel German-technology which was developed to monitor and estimate the tree growth development by looking into the inner structure of the tree to analyse the growth and health status. This technique contributes to detect and study the internal assembly of a tree for the mapping of decay, hollowness, and also to distinguish the sapwood and heartwood demarcation, this way of discovering the internal growth at an early stage, mainly for the timber trees which are economically important can help to regulate thereafter to check whether the growth is not hindered. This paper highlights the device operational methods, electric resistance testing of trees and its applications, also reviewed the various successful application of this equipment in various tree species through-out the world to estimate non-destructively for accurate quantification of standing trees. The performance of this instrument has created breakthrough among various studies and methodologies to spot the internal condition in a standing tree with less invasive ways. PubDate: 2022-05-11
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract According to the existing chaotic image encryption technology, this paper designs an algorithm based on the thought of improved ant colony walking path to encrypt the image. In this paper, the key is generated using SHA-512, and the required chaotic sequence is generated using the PWLCM one-dimensional chaotic system. First, the image is scrambled by the row and column indexes by the improved algorithm so that complete the first scrambling. Then, the image is converted into one-dimensional array, and randomly scrambled according to the designed algorithm. Finally, the one-dimensional array is divided into two segments with the middle point serving as the boundary, and each segment is created using distinct rules based on the array’s number and parity to complete the diffusion, after which the array is restored to its original size. Simulation results show that this algorithm has better performance and robustness in encryption compared with other algorithms. PubDate: 2022-05-03
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Image classification refers to the classification of the input image according to some algorithms. The general steps of image classification include image preprocessing, image feature extraction and image classification judgment. Convolutional neural network (CNN) imitates the visual perception mechanism of biology, solves the complicated engineering of traditional manual feature extraction, and realizes automatic feature extraction from data. However, CNN still has the disadvantages of low efficiency and incomplete feature extraction. In this paper, we propose a depthwise separable squeeze-and-excitation based on multi-feature fusion (DSSEMFF) for image classification. Through feature fusion of multiple models, the network can learn the input features with different levels of images, increase feature complementarity and improve feature extraction ability. By adding the attention module, the network can pay more attention to the targeted area and reduce irrelevant background interference information. Finally, we conduct experiments with other state-of-the-art classification methods, the accuracy is higher than 90% and the error rate is lower than 18% the results show that the effectiveness of the proposed method obtains the better effect. PubDate: 2022-05-03
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract For mobile streaming media service providers, it is necessary to accurately predict the quality of experience (QoE) to formulate appropriate resource allocation and service quality optimization strategies. In this paper, a QoE evaluation model is proposed by considering various influencing factors (IFs), including perceptual video quality, video content characteristics, stalling, quality switching and video genre attribute. Firstly, a no-reference video multimethod assessment fusion (VMAF) model is constructed to measure the perceptual quality of the video by the deep bilinear convolutional neural network. Then, the deep spatio-temporal features of video are extracted using a TSM-ResNet50 network, which incorporates temporal shift module (TSM) with ResNet50, obtaining feature representation of video content characteristics while balancing computational efficiency and expressive ability. Secondly, video genre attribute, which reflects the user’s preference for different types of videos, is considered as a IF while constructing the QoE model. The statistical parameters of other IFs, including the video genre attribute, stalling and quality switching, are combined with VMAF and deep spatio-temporal features of video to form an overall description parameters vector of IFs for formulating the QoE evaluation model. Finally, the mapping relationship model between the parameters vector of IFs and the mean opinion score is established through designing a deep neural network. The proposed QoE evaluation model is validated on two public video datasets: WaterlooSQoE-III and LIVE-NFLX-II. The experimental results show that the proposed model can achieve the state-of-the-art QoE prediction performance. PubDate: 2022-04-18
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract In recent years, the neural networks are frequently adopted to address the issues of cone beam CT imaging. However, most of the research so far has been to build neural networks individually either in the image domain or in the projection domain for specific purposes, while the connections between them are not well exploited. A reconstruction network that focuses on the filtering and backward projection process can fully exploit the potential connections between the projection domain and the image domain, and can be plugged into other learned reconstruction models for high-quality results. However, until now most of the existing reconstruction networks have taken the fully-connected layer as the backbone, which leads to an explosive growth in model parameters, especially in reconstruction of cone-beam CT. In this paper, a sparse-sharing-reconstruction network (SSRNet) with sparse connections and multi-group weight sharing is proposed, which can be regarded as a substitute of the filtering and backward projection process and can significantly reduce the number of network parameters up to 0.4% of that of the previous models. The experimental results show that the reconstruction results of the 2D SSRNet are basically consistent with those of the traditional FBP algorithm. The 3D SSRNet outperforms the traditional FDK algorithm at 50 layers away from the midplane, while still keeping the number of network parameters within a manageable range. PubDate: 2022-03-28
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Cantilever is dramatically used as a resonator sensor to detect the presence of a particular molecule or cell in an environment and to measure their amount. Electrostatic force is commonly used to actuate MEMS based cantilevers to resonate because a cantilever has a simple capacitor structure. In this paper, a novel design is proposed to optimize the cantilevers performance by use of high dielectric material in its capacitor structure. The mathematical model of the proposed design and the performance of the cantilever sensor and its quality has been evaluated and reported by finite element simulations in this way. PubDate: 2022-03-22
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Images captured from extreme surveillance environments may contain missing pixels because of the damaged sensors and dusty particles including debris from combustion fuels and spider webs. This paper addresses the challenging problem of super resolving color images from low resolution images with missing sample values. In conventional methods, missing pixel estimation and super resolution are carried out separately. Unlike the conventional techniques, the proposed method performs simultaneous missing pixel estimation, super resolution and color reconstruction in a unified framework. The problem is posed in a regularized optimization framework with a suitable objective function. The formulated objective function consists of a weighted data fidelity term to control the contribution from each pixel position, a nuclear norm regularization term for enforcing image completion, a three dimensional total variation term for ensuring edge consistency and a sparsity based color regularization term for enforcing proper color reconstruction. Significance of each terms in the regularization function is studied and the results are presented. Experimental results show that proposed method gives improved performance than the conventional methods in most of the cases. PubDate: 2022-03-16
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Image pre-processing is one of the vital tasks used to redefine an image to enhance human visual perception and better information extraction. Several state-of-art have been proposed to de-noise an image contaminated with impulsive noise. This paper focuses on the study and analysis of several de-noising algorithms based on median filter and its advanced non-linear approaches. The study further concentrates on the implementation approaches proposed for de-noising impulsive noise with the deep learning technique. The study highlights the limitation of one approach and its possible solutions with other approaches proposed. The study also highlights several other issues such as optimum selection of window size, edge restoration, even number of noise-free pixels in the median filter, and its variations which have no solution so far. The performance metrics used for the evaluation of several state-of-art algorithms are peak signal-to-noise ratio (PSNR), mean absolute error, computation time and structural similarity index (SSIM). Some of the recently developed algorithms such as classifier and regression model and deep convolutional neural network-based model show PSNR of 45.66 dB and 45.14 dB in 10% noise density and 28.77 dB and 29.18 dB in 90% noise density using Lena image respectively while SSIM of 0.9851 and 0.9847 in 10% noise density and 0.8116 and 0.8101 in 90% noise density using sample1 from BBBC041 dataset respectively. The paper brings out the limitations and issues associated with the conventional and deep learning approaches for the removal of impulsive noise both subjectively and objectively. PubDate: 2022-03-11
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract The purpose of this paper is to present a high-performance pH-ISFET readout circuit, which carries out a temperature insensitivity, linearity and temporal drift compensation, by using a new architecture that automates the control of an isothermal point. Unlike many existing readout circuits in the literature, this circuit can be optimized for several isothermal pH values as desired and for any structure compatible with the standard ISFET sensor. To eliminate the effect of the temporal drift, generally observed in ISFET type sensors, the same readout circuit was used in conjunction with Machine Learning (ML) implementation. The ML model was trained using a dataset from simulations performed using the ISFET macro-model including the drift effect. Through simulations, we show that the proposed scheme reduces drastically the temperature sensitivity of the sensor to less than \(1.5\times 10^{-4}\,{\mathrm{pH}}/^{\circ }{\mathrm{C}}\) for pH \(\pm \,2\) around any isothermal point at a wide pH range (from 1 to 12). For small changes of the pH around the isothermal point, the readout circuit outperforms many other designs with a thermal sensibility of less than \(3.2\times 10^{-6}\,{\mathrm{pH}}/^\circ {\mathrm{C}}\) . Results show that the system was able to predict the long-term behavior of the pH-ISFET (several days) with a relative error, of the output, that not exceed \(0.19\%\) for the 3-sigma testing. PubDate: 2022-03-03 DOI: 10.1007/s11220-022-00378-2
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Vision-based hand gesture recognition involves a visual analysis of handshape, position and/or movement. Most of the previous approaches require complex gesture representation as well as the selection of robust features for proper gesture recognition. To eliminate the problem of illumination variation and occlusion in gesture videos, a simple model-based framework has been presented here using a deep network for hand gesture recognition. The model is fed with ‘hand-trajectory-based-contour-images’. These images represent the motion trajectory of the hand for isolated trajectory gestures obtained via pre-processing steps—a two-level segmentation process and a double-tracking system. Deep features extracted from these images are used for estimating the hand gestures. Conventional machine learning methods involve tedious feature engineering schemes, while deep learning approaches can learn image features hierarchically from local to global with multiple layers of abstraction from a vast number of raw sample images. The feature learning capability of CNN architecture has been used here and it has shown outstanding results on three different datasets. PubDate: 2022-02-22 DOI: 10.1007/s11220-022-00379-1
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Magnetic Resonance Imaging (MRI) is a powerful non-invasive procedure for imaging that offers critical functional, structural, anatomical details about a patient. However, the maximum time needed to scan the whole process causes motion artifacts that may worsen the image quality and result in data distortion and patient discomfort. Hence, an effective mechanism is designed for the reconstruction of MR images. Here, the MR image is gathered from the MRI dataset in the MR image acquisition module. Once the MR image is identified, it is given to the sub-sampling module. After sampling the MR image, it is further given to the weighted compressive sensing module, which is performed using the proposed optimization algorithm, named Renyi Entropy-based Spotted Hyena Fractional Calculus algorithm (RE-SHFC) to get the final reconstructed image. Here, the RE-SHFC is devised by combining Renyi Entropy (RE) measure, and the SHFC algorithm. SHFC is the integration of Spotted Hyena Optimizer and Fractional Calculus. The proposed RE-SHFC algorithm outperformed other methods with minimal Mean Square Error, maximal Peak Signal-to-Noise-ratio, and maximal Structural Similarity Index Module. PubDate: 2022-02-16 DOI: 10.1007/s11220-022-00377-3
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Strain gauge type pressure sensors are widely used in different branches of industry to measure pressure from very low to very high (1400 Mpa) values. This article investigates a strain gauge type pressure sensor that uses silicon oil within its housing to transmit working pressure from the external environment to a sensing plate. An important failure mode arises from loss/leakage of the silicon oil, whereby a portion of the internal volume is replaced by gas, usually air. Coupled nonlinear governing equations have been derived and solved in both static and dynamical states to describe the behavior of the external membrane, the interface oil including pockets of gas, and the sensing plate. Nonlinear behavior arises from the plate and membrane midplane stretching, and of course the behavior of the gas. The resulting model describes how oil loss affects the sensor performance and changes the sensor output and pressure measurable range. PubDate: 2022-01-13 DOI: 10.1007/s11220-021-00376-w
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Segmenting the tongue body is an essential step for automated tongue diagnosis, which is a challenge task due to the tongue body’s specificity and heterogeneity. The current deep-learning based tongue image segmentation networks are bloated with high computational complexity. In this study, a light-weight segmentation network for tongue images is proposed under the basic encoder-decoder framework, in which MobileNet v2 is adopted as the backbone network, due to its few parameters and low computational complexity. The high-level semantic information and low-level positional information are combined together to detect the tongue body’s boundary. And the dilated convolution operations are performed on the final feature maps of the network to enlarge the receptive field, so as to capture rich global semantic information. An attention mechanism is embedded to re-calibrate the feature maps spatially and channel-wise to enhance important features for the segmentation task, while suppressing the irrelevant ones. Moreover, a supervision output is added to each level of the decoder to guide the network to capture both the local and global image features for accurate tongue image segmentation. All supervision outputs are fused to produce good segmented results. The quantitative and qualitative results on two tongue datasets indicate that the proposed network can achieve a competitive performance with smaller model size and lower computational cost. The proposed method could accurately extract the tongue body, which can fully meet the requirements of practical applications. PubDate: 2022-01-08 DOI: 10.1007/s11220-021-00375-x
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Electromagnetic tomography (EMT) is an emerging imaging modality capable of visualizing the distribution of electrically conductive or magnetically permeable materials within the vessels and pipelines. Image reconstruction is the crucial step of EMT inverse problem, which is one of the main challenges in the promotion and application of EMT to industrial and biomedical fields. EMT inverse problem is ill-posed and ill-conditioned, which is the key to the reconstructed images quality. Tikhonov regularization is the widely used regularization method to solve the ill-posed problem. The revised Tikhonov regularization is obtained by improving the Tikhonov regularization when observation noise is considered. This paper presents an improved conjugate gradient algorithm based on the revised Tikhonov regularization to reduce the ill-posed nature of EMT inverse problem and enhance the spatial resolution of the reconstructed images. Numerical simulations results confirm that the quality of the reconstructed images obtained by the proposed algorithm is improved and also better than the other conventional algorithms including linear back projection, Tikhonov regularization algorithm, Landweber iterative algorithm, in terms of the typical patterns. Besides, correlation coefficients of the proposed ICGRT algorithm are 0.5961, 0.5989 and 0.5231 for three experimental typical patterns. The proposed ICGRT algorithm has highest correlation coefficients and reconstructs the best images. PubDate: 2022-01-07 DOI: 10.1007/s11220-021-00374-y
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract This paper proposes a self-reset pulse frequency modulation (PFM) digital pixel sensor (DPS) with in-pixel variable reference voltage for optical brain imaging systems. The sensor demonstrates a wide dynamic range and very low power consumption that can detect small signals of brain activity in brain. The high dynamic range, high SNR (signal-to-noise ratio), high speed and low power consumption image sensor are suitable for optical brain imaging systems. Since the comparator part consumes high power inside pixel, sub-threshold, self-biased and bulk-driven techniques are used to achieve both ultra-low-voltage and low power in the PFM DPS. Moreover, High speed (high frame rate) is achieved by image capturing in-parallel for all pixels. The proposed image sensor is post layout simulated in 0.18 µm Complementary Metal Oxide Semiconductor (CMOS) technology with 0.6 V supply voltage, resulting in the dynamic range of 152 dB and the power consumption of 11.25 nW and the fill factor of the proposed sensor is 11%. Hence, this device has significant potential to be used for brain signal detection in pre-clinical and clinical studies, cognitive process, diagnose diseases in exploring brain structure and function. PubDate: 2021-12-25 DOI: 10.1007/s11220-021-00373-z
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract The rising rates of costs in labor and growth of computerization in fabric industries made defect detection in fabric a promising domain. For a huge time, manual discovery is extensively utilized in textile industries by trained staff that results in high cost. Meanwhile, the strict quality assessment is done by modern textile industries, which made automatic fabric defect detection a reliable choice. Since defect detection is an important and challenging aspect of modern industrial manufacturing, it is necessary to determine the quality and acceptability of garments and to reduce the cost and time waste caused by defects. Different methods are in practice for effective detection of fabric defects, however, they limit due to many reasons. Thus we proposed a new method named Competitive Cat Swarm Optimizer (CCSO) based Deep neuro-fuzzy network (DNFN) for effective fabric detection. Here, the pre-processing is performed with a median filter for eliminating noise contained in the image. Furthermore, features are extracted that involves Tetrolet transform-based features, statistical features, like energy, entropy, homogeneity, contrast, correlation, and texture feature, like Local gradient pattern. The data augmentation is carried out based on obtained features to make it apposite for processing. The defect detection is carried out using RideNN and DNFN. Here, the training of DNFN is done using the proposed CCSO, which is devised by combining Cat Swarm Optimization and Competitive Swarm Optimizer. The correlation is performed on the outputs of RideNN and DNFN for generating the final output of fabric defect detection. The performance of the proposed CCSO-based DNFN is compared with the different existing methods and the proposed CCSO-based DNFN outperforms with the highest specificity of 0.920, the accuracy of 0.919, and sensitivity of 0.916. PubDate: 2021-12-15 DOI: 10.1007/s11220-021-00370-2
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract The main objective of this article is to analyse strategies of embodied cognition and the intersubjective ground for individual intentions in the process of image-based oncoradiology diagnosis. The article presents a range of both oncoradiology imaging specifics and concrete operations performed by radiologists during their daily professional routine. This data shows how the embodied diagnostic cognition based on medical imaging is structured. Hence, this paper proposes an enactive theory of oncoradiology imaging and considers the wider problem of how knowledge is related to the (embodied) subjectivity in a particular social setting. PubDate: 2021-12-04 DOI: 10.1007/s11220-021-00372-0
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract The nanostructured β-Bi2O3 thin film was deposited on glass substrates by chemical spray pyrolysis technique using the mixture of bismuth nitrate pentahydrate with deionized water and nitric acid as a precursor solution. The thin film deposition condition and the precursor salt concentration were optimized to obtain nanostructured β-Bi2O3 thin films. The film obtained from 0.05 M of bismuth nitrate pentahydrate aqueous solution was sprayed at the rate of 3 mL/min. on pre-heated glass substrate at the temperature of 250 °C yielded spherical shaped well-connected nanocrystallites, which has large surface area. The diffraction peak position in XRD confirmed the formation of crystalline β-Bi2O3 with tetragonal crystal structure. Further sensing characteristics of β-Bi2O3 thin film towards various dimethylamine (DMA) vapour concentration have been investigated. The sensing results revealed that β-Bi2O3 thin film shows good sensing response towards dimethylamine vapour at an ambient temperature. The minimum detection limit was found to be 0.5 ppm, and sensors show shorter response and recovery time (28 s and 10 s). The dimethylamine sensing characteristics (response, sensitivity, electrical hysteresis, selectivity in mixed vapour environment, stability) of β-Bi2O3 thin films were discussed and reported. PubDate: 2021-12-03 DOI: 10.1007/s11220-021-00371-1
Please help us test our new pre-print finding feature by giving the pre-print link a rating. A 5 star rating indicates the linked pre-print has the exact same content as the published article.
Abstract: Abstract Sarcosine is one of the known small molecule biomarkers to detect prostate cancer effectively. In the present work, we presented a very simple procedure and low-cost non-enzymatic method for the detection of Sarcosine. Zinc oxide (ZnO) nanoparticles were successfully synthesized, characterized and elucidated the morphology on the Indium tin oxide (ITO) surface, which showed uniform arrangement with a spherical shape. The ITO working electrode modified with ZnO exhibits better analytical characteristics for Sarcosine sensing with a linear range between 5 and 100 nM. The limit of detection was found to be low (7.5 nM) with excellent sensitivity and possess quick response time. Due to its high specificity and repeatability, the ITO/ZnO working electrode does not interfere with the other amino acids in the real samples. PubDate: 2021-11-13 DOI: 10.1007/s11220-021-00369-9
Hybrid journal (It can contain Open Access articles)
