Shape Texture Research Articles

Quantifying Shape and Texture Biases for Enhancing Transfer Learning in Convolutional Neural Networks

Neural networks have inductive biases owing to the assumptions associated with the selected learning algorithm, datasets, and network structure. Specifically, convolutional neural networks (CNNs) are known for their tendency to exhibit textural biases. This bias is closely related to image classification accuracy. Aligning the model’s bias with the dataset’s bias can significantly enhance performance in transfer learning, leading to more efficient learning. This study aims to quantitatively demonstrate that increasing shape bias within the network by varying kernel sizes and dilation rates improves accuracy on shape-dominant data and enables efficient learning with less data. Furthermore, we propose a novel method for quantitatively evaluating the balance between texture bias and shape bias. This method enables efficient learning by aligning the biases of the transfer learning dataset with those of the model. Systematically adjusting these biases allows CNNs to better fit data with specific biases. Compared to the original model, an accuracy improvement of up to 9.9% was observed. Our findings underscore the critical role of bias adjustment in CNN design, contributing to developing more efficient and effective image classification models.

Effects of texture arrangement and texture shape on the static characteristics of a four-lobe bearing

Effects of texture arrangement and texture shape on the static characteristics of a four-lobe bearing

Comparative Analysis of Classification and Segmentation Performance of Different Dog Breeds Based on Mask R-CNN

The classification of different image data is a common task in real life, which has been achieved as early as convolutional neural networks. With the emergence of R-CNN (Region CNN), simple classification tasks began to develop towards object detection tasks, followed by various segmentation tasks. After Faster R-CNN [4], Mask R-CNN [3] achieved instance segmentation while detecting objects. This algorithm model needs to distinguish different target objects in the image panel and learn a large number of characteristics to represent the details of each target object. In dog data images, there are differences and similarities in the characteristics of dog species, such as fur color, texture, and body shape. So in this article, the author selected and created two different dog datasets, representing targets with significant differences and similarities in features. Compare and analyze the performance of Mask R-CNN object detection and instance segmentation using these two dog datasets.

Unveiling the reinforcement benefits of innovative textured geogrids

The smooth surface texture of the commercially available geogrids limits the shear strength mobilization at the interfaces. This study presents the design, manufacturing, and interface performance evaluation of innovative textured geogrids. Geogrids with square, triangular, and hexagonal apertures with and without inherent surface texture were manufactured through additive manufacturing (3D printing) technique, using PLA (Poly Lactic Acid) filament. The texture includes elevated pins of 3 mm height at the junctions and inherent diamond pattern of 1 mm height on the ribs. The individual and combined effects of surface texture and aperture shape on the stress–displacement relationship, dilation angle, and the thickness of shear zone are quantified using large-scale direct shear tests and Particle Image Velocimetry (PIV) analysis. Results showed that the textured geogrid with hexagonal aperture has exhibited the maximum interface coefficient of 0.96 with sand followed by the geogrids with triangular and square apertures. Irrespective of the aperture shape, provision of the surface texture resulted in an overall increase of interface shear strength by more than 13%. Further, PIV analysis revealed that the shear zone is 25% thicker for textured geogrids of different aperture shapes, suggesting higher interlocking and passive resistance offered by their textured surfaces.

Links between soil pore structure, water flow and solute transport in the topsoil of an arable field: Does soil organic carbon matter?

An improved understanding of preferential solute transport in soil macropores would enable more reliable predictions of the fate of agrochemicals and the protection of water quality in agricultural landscapes. The objective of this study was to investigate how soil organic carbon (SOC) and soil texture shape soil pore structure and thereby determine the susceptibility to preferential transport under steady-state near-saturated flow conditions. To do so, we took intact topsoil samples from an arable field that has large variations in SOC content (1.1–2.7%) and clay content (8–42%). Soil pore structure was quantified by X-ray tomography and soil water retention measurements. Non-reactive solute transport experiments under steady-state near-saturated conditions were carried out at irrigation rates of 2 and 5 mm h−1 to quantify the degree of preferential transport. Near-saturated hydraulic conductivities at pressure heads of −1.3 and −6 cm were also measured using a tension disc infiltrometer. The results showed that larger abundances of small macropores (240–720 µm diameter) and mesopores (5–100 µm diameter) resulted in weaker preferential transport, due to larger hydraulic conductivities in the soil matrix that prevented the activation of water flow and solute transport in large macropores. In particular, the degree of preferential transport was most strongly and negatively correlated with the mesoporosity in the 30–100 µm diameter class. In contrast, the degree of preferential transport was not correlated with connectivity measures (e.g. the percolating fraction and critical pore diameter for the macropore network), probably because i.) the pore space of almost all samples was highly connected, being dominated by one percolating cluster, and ii.) only a part of this percolating macroporosity was active under the near-saturated conditions of the experiment. We also found that the degree of preferential transport was strongly and negatively correlated with clay content, whilst the effects of SOC were not significant. Nevertheless, macroporosity in the 240–720 µm diameter class and mesoporosity were positively correlated with SOC content in our soils and in some previous studies. Therefore, SOC sequestration in arable soils may potentially reduce the risk of preferential transport under near-saturated flow conditions through better developed networks of small macropores and mesopores.

Optimum texture shape under different lubrication conditions applied to the start-up phase of journal bearings

This paper presents a parameterized model to optimize the texture units using an improved particle swarm optimization. The texture units are optimized under three lubrication states: boundary lubrication, mixed lubrication, and hydrodynamic lubrication, by utilizing the Stribeck curve. The optimal texture units can be applied to journal bearings and thrust bearings in different operating conditions to achieve maximum load-carrying capacity. The paper applies the optimal textures under the three lubrication states to the start-up stage of journal bearings to examine the lubrication performance and dynamic characteristics of the textured journal bearing system. The results demonstrate that compared to the untextured journal-bearing system, the introduction of optimal textures improves the tribological performance of the journal-bearing system, and the optimal textured units under hydrodynamic lubrication have the best friction reduction effect. Furthermore, it is found that the introduction of textures has no significant effect on the dynamic characteristics of the journal bearings.

A Study on the Steady and Dynamic Performance of Dry Gas Seals with Combined Superelliptical Grooves and Holes on the End Face

Background:: Enhancing the stability and leakage control of Dry Gas Seals (DGS) under high parameters has been a crucial research focus. The design of end-face groove structures and surface texture shapes has been an essential aspect of DGS studies aimed at improving performance. Objective:: The proposed end-face gas seal utilizes superelliptical grooves and holes to improve its performance, aiming to obtain a patent. The use of superelliptical curves allows for a more precise and efficient geometric representation, resulting in a better sealing effect. Methods:: A theoretical analysis model for the steady and dynamic behavior of the seal is established using the lubrication theory and perturbation methods. The study investigates the distribution patterns of gas film pressure on the sealing end-face and gas flow characteristics within the film. This approach provides a new perspective for understanding seal performance and offers a theoretical basis for optimizing seal design. Results:: The results indicate that the combined end-face structure with grooves and holes ensures good sealing stability and effectively enhances leakage control performance. By optimizing the design of the superelliptical groove and holes on the end face, the performance of DGS can be significantly improved. Conclusion:: Within the parameter range studied, better steady-state performance is achieved for θ=40~80°, v=1.3~1.4, u=1~2, β=0.6~0.7, and λ=1.0~1.5. In addition, better dynamic performance is observed for θ=80~120°, v=1.1~1.2, u=2~3, β=0.9~1.0, and λ=2.0~2.5.

Influence of femtosecond laser textured shapes on the friction characteristics of 7075 aluminum alloy

Different surface dimples (circular, square, and triangular) were fabricated on 7075 aluminum alloy specimens using a femtosecond laser to enhance their wear resistance. The influence of these surface dimples on the friction performance of the aluminum alloy under dry sliding conditions was investigated by comparing the coefficient of friction (COF), wear track morphology, and wear rate. All the textured samples exhibit varying degrees of improvement in the wear resistance compared to the untextured sample, and the circular texture is the most effective. Wear track analysis reveals that abrasive and adhesive wear are the primarily wear mechanisms. Thus, introducing surface dimples significantly enhances the wear resistance of the 7075 aluminum alloy, providing a novel strategy for optimizing its friction performance through design.

Numerical study on tribology properties of textured surfaces based on a new contact model

The frictional properties of textured surfaces are closely related to the contact between rough peaks in the mixed lubrication. A model for rough peak contact between polyurethane and metal was already developed in a previous study. This study aims to further analyze the influence of various parameters on the friction properties of textured surfaces based on the previously developed theoretical model. As found, the friction coefficient increases with the increase of the root mean square roughness, and the increasing magnitude shows an upward trend. The friction coefficient varies with different texture shapes, and among the four shapes, the spherical texture has the smallest friction coefficient. In addition, either small or large texture depths can adversely affect the friction coefficient. The friction coefficient decreases and stabilizes with the increase of area ratio. The effect of working parameters on the friction coefficient coincides with the Stribeck curve.

Contributions to the taxonomy of Periplocoideae and Asclepiadoideae (Apocynaceae) in Türkiye based on fruit and seed morphology

In this study, the fruit and seed morphology of all the Periplocoideae (Periploca gracilis and P. graeca) and the majority of Asclepiadoideae (Araujia sericifera, Cionura erecta, Cynanchum acutum and Gomphocarpus fruticosus) taxa in Türkiye was evaluated from a taxonomic point of view using, for the first time, both light and scanning electron microscopy. Macro- and micro-morphological investigations were carried out on herbarium samples. An identification key for the examined Periplocoideae and Asclepiadoideae taxa, and expanded fruit and seed descriptions for each taxon were prepared. Numerical analysis showed that shape and size of follicle, size, shape and surface texture of seed were found to be effective in delimiting the studied Periplocoideae and Asclepiadoideae taxa at subfamilial and generic levels.

SHOWMe: Robust object-agnostic hand-object 3D reconstruction from RGB video

In this paper, we tackle the problem of detailed hand-object 3D reconstruction from monocular video with unknown objects, for applications where the required accuracy and level of detail is important, e.g. object hand-over in human–robot collaboration, or manipulation and contact point analysis. While the recent literature on this topic is promising, the accuracy and generalization abilities of existing methods are still lacking. This is due to several limitations, such as the assumption of known object class or model for a small number of instances, or over-reliance on off-the-shelf keypoint and structure-from-motion methods for object-relative viewpoint estimation, prone to complete failure with previously unobserved, poorly textured objects or hand-object occlusions. To address previous method shortcomings, we present a 2-stage pipeline superseding state-of-the-art (SotA) performance on several metrics. First, we robustly retrieve viewpoints relying on a learned pairwise camera pose estimator trainable with a low data regime, followed by a globalized Shonan pose averaging. Second, we simultaneously estimate detailed 3D hand-object shapes and refine camera poses using a differential renderer-based optimizer. To better assess the out-of-distribution abilities of existing methods, and to showcase our methodological contributions, we introduce the new SHOWMe benchmark dataset with 96 sequences annotated with poses, millimetric textured 3D shape scans, and parametric hand models, introducing new object and hand diversity. Remarkably, we show that our method is able to reconstruct 100% of these sequences as opposed to SotA Structure-from-Motion (SfM) or hand-keypoint-based pipelines, and obtains reconstructions of equivalent or better precision when existing methods do succeed in providing a result. We hope these contributions lead to further research under harder input assumptions. The dataset can be downloaded at https://download.europe.naverlabs.com/showme.

Numerical Optimization Analysis of Floating Ring Seal Performance Based on Surface Texture

Much research and practical experience have shown that the utilization of textures has an enhancing effect on the performance of dynamic seals and the dynamic pressure lubrication of gas bearings. In order to optimize the performance of floating ring seals, this study systematically analyzes the effects of different texture shapes and their parameters. The Reynolds equation of the gas is solved by the successive over-relaxation (SOR) iteration method. The pressure and thickness distributions of the seal gas film are solved to derive the floating force, end leakage, friction, and the ratio of buoyancy to leakage within the seal. The effects of various texture shapes, including square, 2:1 rectangle, triangle, hexagon, and circle, as well as their parameters, such as texture depth, angle, and area share, on the sealing performance are discussed. Results show that the texture can increase the air film buoyancy and reduce friction, but it also increases the leakage by a small amount. Square textures and rectangular textures are relatively effective. The deeper the depth of the texture within a certain range, the better the overall performance of the floating ring seal. As the texture area percentage increases, leakage tends to increase and friction tends to decrease. A fractal roughness model is developed, the effect of surface roughness on sealing performance is briefly discussed, and finally the effect of surface texture with roughness is analyzed. Some texture parameters that can significantly optimize the sealing performance are obtained. Rectangular textures with certain parameters enhance the buoyancy of the air film by 81.2%, which is the most significant enhancement effect. This rectangular texture reduces friction by 25.8% but increases leakage by 79.5%. The triangular textures increase buoyancy by 28.02% and leakage increases by only 10.08% when the rotation speed is 15,000 r/min. The results show that texture with appropriate roughness significantly optimizes the performance of the floating ring seal.

Dynamic and stability analysis of crescent geometry-possessing textured journal bearing using nanolubricant

PurposeThis study aims to find the dynamic performance parameters of the journal bearing with micro geometries patterning the arc (crescent) shape textures provided in three specific regions of the journal bearing: the full, the second half and the increasing pressure region. The dynamic behavior of textured journal bearings has been analyzed by computing dynamic parameters and linear and non-linear trajectories.Design/methodology/approachThe lubricant flows between the bearing and journal surface are governed by Reynold’s equation, which has been solved by finite the element method. The dynamic performance parameters such as stiffness, damping, threshold speed, critical mass and whirl frequency ratio are examined under various operating conditions by considering various ranges of eccentricity ratios and texture depths. Linear and non-linear equations of motion have been solved with Ranga–Kutta method to get journal motion trajectories. Also, the impact of adding aluminum oxide and copper oxide nanoparticles to the base lubricant in combination with arc-shaped textures is analyzed to further see any enhancement in the performance parameters.FindingsThe findings demonstrated that direct stiffness and damping parameters increased to their maximum level with six textures in the pressure-increasing region when compared with the untextured surface. Also, nanoparticle additives showed improvements above the highest value attained with no inclusion of additives in the same region or quantity of textures.Originality/valueEngineers may design bearings with improved stability and overall performance if they understand how texture form impacts dynamic properties.

A geometric approach to robust medical image segmentation

Robustness of deep learning segmentation models is crucial for their safe incorporation into clinical practice. However, these models can falter when faced with distributional changes. This challenge is evident in magnetic resonance imaging (MRI) scans due to the diverse acquisition protocols across various domains, leading to differences in image characteristics such as textural appearances. We posit that the restricted anatomical differences between subjects could be harnessed to refine the latent space into a set of shape components. The learned set then aims to encompass the relevant anatomical shape variation found within the patient population.We explore this by utilising multiple MRI sequences to learn texture invariant and shape equivariant features which are used to construct a shape dictionary using vector quantisation. We investigate shape equivariance to a number of different types of groups. We hypothesise and prove that the greater the group order, i.e., the denser the constraint, the better becomes the model robustness. We achieve shape equivariance either with a contrastive based approach or by imposing equivariant constraints on the convolutional kernels. The resulting shape equivariant dictionary is then sampled to compose the segmentation output. Our method achieves state-of-the-art performance for the task of single domain generalisation for prostate and cardiac MRI segmentation. Code is available at https://github.com/AinkaranSanthi/A_Geometric_Perspective_For_Robust_Segmentation.

IMG-13. PREDICTION OF PEDIATRIC MEDULLOBLASTOMA SUBGROUPS USING CLINICO-RADIOMIC ANALYSIS

Abstract BACKGROUND Pediatric medulloblastoma is an aggressive brain tumor and tailored treatment has the potential to lead to better patient outcomes. The WHO 2021 classification of medulloblastoma involves an integrated diagnosis that incorporates genetically defined characteristics, including molecular subgroups (WNT-activated, SHH-activated TP53 wildtype, SHH-activated TP53-mutant, and non-WNT/non-SHH). Clinically acquired radiology (MRI) imaging characteristics could provide a non-invasive, pre-treatment biomarker for such subgroups faster than standard methylation-based turn-around times, and for use in countries where methylation is not available. METHODS Herein, we utilize a multi-institutional dataset of multi-parametric, clinical MRIs of pediatric medulloblastoma patients from the Children’s Brain Tumor Network (median age = 8.9 years) to assess the predictive value of patient-level radiomic and clinical factors. 89 subjects with treatment-naïve T1w/T1w contrast-enhanced/T2w/FLAIR images and molecular subgroups (derived from methylation profiling or RNA sequencing) were included (SHH=18, WNT=10, non-SHH/WNT=61). Radiomic features were extracted from a radiologist-defined volumetric (3D) segmentation for each subject separately (including solid tumor, cystic, and peritumoral edema regions). Clinical variables included sex, age at diagnosis, and metastatic disease. 20 top performing features were selected based on an ANOVA between features/classes and subsequently used to train and evaluate three separate classification models based on radiomic, clinical, or clinico-radiomic features (Linear SVM; leave-one-subject-out cross-validation). RESULTS The combined clinico-radiomic model had the top performance (AUC=0.83) followed by radiomic (0.78) and clinical (0.5) for predicting SHH/WNT vs. non-SHH/WNT groups. Predictive radiomic features included intensity-based statistics (T1w, T2w, T1w-CE), texture (T2w), and morphological shape characteristics. CONCLUSIONS Our findings show potential for early prediction of molecular subgroups using baseline imaging that could lend to faster decision-making in patient treatment planning. Future work aims to evaluate the inclusion of histological characteristics for a full integrated diagnostic approach as well as assessment of prognostic value of the combined multi-omic features.

Effect of textured shapes on slot entry hybrid conical journal bearing with ER lubricant behavior

This article investigates the influence of textured shapes and electrorheological fluid (ERF) behavior on slot entry hybrid conical journal bearing system (HCJBs). The slot entry restrictor is used as a compensating element to supply the lubricant in the clearance space. Spherical, rectangular, and triangular textured shapes over the surface of conical bearings have been employed. Furthermore, the intelligent properties of ERF on the behavior of textured slot entry HCJBs have been analyzed. The finite element method (FEM) and the successive iteration-based Gauss-Seidel method have been used to solve the modified Reynolds equation. The results drawn from this study show different influences on the bearing performance for different textured shapes. Using textured surfaces and ERF improves the minimum fluid film thickness, stability, and bearing stiffness compared to the non-textured bearing.

Enhancement mechanism of asphalt mixture skeleton structures due to morphological characteristics of steel slag

Uncertainty still exists about the enhancement mechanism of the skeleton structure brought about by the steel slag, and the relationship between skeleton structure of steel slag asphalt mixture and the mechanical properties is still ambiguous. This study quantified the skeleton structure of steel slag asphalt mixture and explored the enhancement mechanism of skeleton structure brought about by the steel slag. The relationship between skeleton parameters and mechanical properties of steel slag asphalt mixture was also established. The results show that recycling steel slag as aggregate improves the skeleton quality of mix. The slag particles are not susceptible to be oriented tending to the horizontal direction. The abundant angularity of steel slag increased the contact number while the rough texture and non-spherical shape of steel slag enhanced the contact length. This project provided new insight and a notable opportunity to advance the comprehension of the skeleton structure of asphalt mixture containing steel slag.

Performance of Sustainable Green Concrete Incorporating Quarry Dust and Ferronickel Slag as Fine Aggregate.

This paper presents a study on the combined use of two by-products, namely quarry dust (QD) and ferronickel slag (FNS), as a full substitute for natural sand to improve the greenness of concrete production. Quarry dust was used in increments of 25% to a maximum of 75% substitution, where nickel slag was used as the remaining proportion of fine aggregate. All the combinations of quarry dust and nickel slag were found to be compliant with AS 2758.1 and they showed notably better grading than 100% sand. In this research, standard concrete tests, such as the slump test for fresh concrete, and compression, tensile and shrinkage tests for hardened concrete, were conducted. Scanning electron microscopy and X-ray diffraction analysis were also conducted for microstructural investigation. The results concluded that the combinations of quarry dust and nickel slag in concrete as a whole substitution of sand provide similar results for these properties. Specifically, 25% quarry dust with 75% nickel slag proved to be the most promising alternative to sand, with compressive and splitting tensile strengths of 62 and 4.29 MPa, respectively, which were 16% and 20% higher than those of the control mix. Also, lower drying shrinkage was observed for this combination compared to the control mix. The higher strength is attributed to the rough texture and angular shape of both quarry dust and nickel slag providing a better mechanical interlocking. The validity of this result has also been confirmed through image analysis of micrographs from various specimens. In microstructural investigations, specimens with QD and FNS exhibited fewer voids and a more compact surface compared to the control specimen. This shows the potential for further research into the use of quarry dust and nickel slag in the production of green concrete.

The determination of heterogeneous catalytic performance of Palm Oil Fuel Ash (POFA) in the production of biodiesel via transesterification of Jatropha oil

The high energy demand with the negative environmental impacts of using fossil fuel for energy generations raised question on over dependability on it for sustainable economic growth. Biodiesel tend to be one of the renewable alternative solutions towards the above problems. Biodiesel can be produced through various methods such as transesterification, micro emulsion and pyrolysis The heterogeneous catalytic performance of Palm Oil Fuel Ash (POFA) in the production of biodiesel via transesterification of jatropha oil was investigated. The morphological and chemical properties of POFA were studied using Scanning Electron Microscopy (SEM) and X-ray Fluoroscopy (XRF) respectively. The crude Jatropha oil was transesterified and 0.1 wt%, 0.2 wt%, 0.3wt%, 0.4wt% and 0.5wt% of POFA were used as heterogeneous catalyst. Fourier Transform Infrared (FTIR) has been used to determine the functional group of the samples. SEM indicates the presence porous structure, texture and irregular shape on POFA while XRF shows that it comprises mainly of SiO2 (79.45 %). The maximum percentage of biodiesel yield was 92.30% at the application of 0.2wt% POFA. This shows that POFA can be used as a heterogeneous catalyst in the production of biodiesel.

Research on tribological performance of textured thrust bearing using gray relational degree and improved multi-objective water circulation algorithm

To enhance the loading capacity of the oil film, reduce the lubricating oil temperature, and minimize the deformation of the friction pairs, three types of concave textures were proposed. The tribological properties of the thrust bearing underwent multi-objective optimization analysis using the two different methods. By comparing the two methods, it was found that the improved algorithm (MOWCA) demonstrated better optimization performance. Under this algorithm, when the texture shape was triangular, the number of textures was 18n, the size was approximately 0.135 mm, and the spacing between textures was approximately 4.6 mm, the loading capacity of the oil film increased by approximately 26.8%, and the temperature decreased by approximately 0.46%.