Shape Of Object Research Articles

Generating implicit object fragment datasets for machine learning

One of the primary challenges inherent in utilizing deep learning models is the scarcity and accessibility hurdles associated with acquiring datasets of sufficient size to facilitate effective training of these networks. This is particularly significant in object detection, shape completion, and fracture assembly. Instead of scanning a large number of real-world fragments, it is possible to generate massive datasets with synthetic pieces. However, realistic fragmentation is computationally intensive in the preparation (e.g., pre-factured models) and generation. Otherwise, simpler algorithms such as Voronoi diagrams provide faster processing speeds at the expense of compromising realism. In this context, it is required to balance computational efficiency and realism. This paper introduces a GPU-based framework for the massive generation of voxelized fragments derived from high-resolution 3D models, specifically prepared for their utilization as training sets for machine learning models. This rapid pipeline enables controlling how many pieces are produced, their dispersion and the appearance of subtle effects such as erosion. We have tested our pipeline with an archaeological dataset, producing more than 1M fragmented pieces from 1,052 Iberian vessels (). Although this work primarily intends to provide pieces as implicit data represented by voxels, triangle meshes and point clouds can also be inferred from the initial implicit representation. To underscore the unparalleled benefits of CPU and GPU acceleration in generating vast datasets, we compared against a realistic fragment generator that highlights the potential of our approach, both in terms of applicability and processing time. We also demonstrate the synergies between our pipeline and realistic simulators, which frequently cannot select the number and size of resulting pieces. To this end, a deep learning model was trained over realistic fragments and our dataset, showcasing similar results.

Non-native language comprehenders encode implied shapes of objects in memory

Barsalou (1999) proposes that conceptual knowledge is represented by mental simulations containing perceptual information derived from actual experiences. Although a substantial number of studies have provided evidence consistent with this view in native language comprehension, it remains unclear whether the non-native language comprehension processes also include mental simulations. The current study successfully replicates the shape match effect in sentence-picture verification (Zwaan et al., 2002) for non-native English language comprehenders, indicating native-like visual simulations. In addition, participants displayed better delayed recognition memory when the shape of the depicted objects matched the shape that was implied by the sentence than when it did not, suggesting that visual simulations were generated spontaneously in naturalistic non-native language comprehension. Additional correlational analyses revealed no relationship between English proficiency and the size of the match effect.

Speckle autocorrelation separation method based on frequency intensity for tracking moving object beyond optical memory effect

Abstract Imaging objects behind a scattering medium beyond the optical memory effect (OME) range is vital in various imaging scenarios. Despite numerous proposed imaging methods aimed at separating objects, effectively tracking moving objects remains challenging. In this paper, we present a non-invasive scattering imaging method to reconstruct target beyond the OME range and track moving objects using a speckle separation and position estimation technique. A speckle autocorrelation separation method based on frequency intensity is proposed to eliminate the noise from the cross-correlation and to separate the objects from the mixed speckles. A position estimation technique is proposed to determine the positional relationships of the objects using speckle deconvolution using the point spread function of imaging regions. A non-invasive imaging system is designed to track the moving object without prior knowledge of the shape and distribution of objects. Experimental validation shows the efficiency of the method for separation and localization, allowing for the reconstruction of the target behind the scattering medium and tracking of the moving object behind the scattering medium. Moreover, the method is capable of separating and tracking a moving object when multiple objects are in the background.

Evaluation of Haptic Textures for Tangible Interfaces for the Tactile Internet

Every texture in the real world provides us with the essential information to identify the physical characteristics of real objects. In addition to sight, humans use the sense of touch to explore their environment. Through haptic interaction we obtain unique and distinct information about the texture and the shape of objects. In this paper, we enhance X3D 3D graphics files with haptic features to create 3D objects with haptic feedback. We propose haptic attributes such as static and dynamic friction, stiffness, and maximum altitude that provide the optimal user experience in a virtual haptic environment. After numerous optimization attempts on the haptic textures, we propose various haptic geometrical textures for creating a virtual 3D haptic environment for the tactile Internet. These tangible geometrical textures can be attached to any geometric shape, enhancing the haptic sense. We conducted a study of user interaction with a virtual environment consisting of 3D objects enhanced with haptic textures to evaluate performance and user experience. The goal is to evaluate the realism and recognition accuracy of each generated texture. The findings of the study aid visually impaired individuals to better understand their physical environment, using haptic devices in conjunction with the enhanced haptic textures.

Analytical form of the refocused images from correlation plenoptic imaging

Correlation plenoptic imaging (CPI) is emerging as a promising approach to light-field imaging (LFI), a technique for concurrently measuring light intensity distribution and propagation direction of light rays from a 3D scene. LFI thus enables single-shot 3D imaging, offering rapid volumetric reconstruction. The optical performance of traditional LFI, however, is limited by a micro-lens array, causing a decline in resolution as 3D capabilities improve. CPI overcomes these limitation by measuring photon number correlations on two photodetectors with spatial resolution, in a lenslet-free design, so that the correlation function can be decoded in post-processing to reconstruct high-resolution images. In this paper, we derive the analytical expression of CPI images reconstructed through refocusing, addressing the previously unknown mathematical relationship between object shape and its final image. We show that refocused images are not limited by numerical aperture-induced blurring, as in conventional imaging. Rather, the image features of CPI can be explained through an analogy with imaging systems illuminated by spatially coherent light.

Augmenting perceived stickiness of physical objects through tactile feedback after finger lift-off.

Haptic Augmented Reality (HAR) is a method that actively modulates the perceived haptics of physical objects by presenting additional haptic feedback using a haptic display. However, most of the proposed HAR research focuses on modifying the hardness, softness, roughness, smoothness, friction, and surface shape of physical objects. In this paper, we propose an approach to augment the perceived stickiness of a physical object by presenting additional tactile feedback at a particular time after the finger lifts off from the physical object using a thin and soft tactile display suitable for HAR. To demonstrate this concept, we constructed a thin and soft tactile display using a Dielectric Elastomer Actuator suitable for HAR. We then conducted two experiments to validate the effectiveness of the proposed approach. In Experiment 1, we showed that the developed tactile display can augment the perceived stickiness of physical objects by presenting additional tactile feedback at appropriate times. In Experiment 2, we investigated the stickiness experience obtained by our proposed approach and showed that the realism of the stickiness experience and the harmony between the physical object and the additional tactile feedback are affected by the frequency and presentation timing of the tactile feedback. Our proposed approach is expected to contribute to the development of new applications not only in HAR, but also in Virtual Reality, Mixed Reality, and other domains using haptic displays.

YOLOv8-Coal: a coal-rock image recognition method based on improved YOLOv8.

To address issues such as misdetection and omission due to low light, image defocus, and worker occlusion in coal-rock image recognition, a new method called YOLOv8-Coal, based on YOLOv8, is introduced to enhance recognition accuracy and processing speed. The Deformable Convolution Network version 3 enhances object feature extraction by adjusting sampling positions with offsets and aligning them closely with the object's shape. The Polarized Self-Attention module in the feature fusion network emphasizes crucial features and suppresses unnecessary information to minimize irrelevant factors. Additionally, the lightweight C2fGhost module combines the strengths of GhostNet and the C2f module, further decreasing model parameters and computational load. The empirical findings indicate that YOLOv8-Coal has achieved substantial enhancements in all metrics on the coal rock image dataset. More precisely, the values for AP50, AP50:95, and AR50:95 were improved to 77.7%, 62.8%, and 75.0% respectively. In addition, optimal localization recall precision (oLRP) were decreased to 45.6%. In addition, the model parameters were decreased to 2.59M and the FLOPs were reduced to 6.9G. Finally, the size of the model weight file is a mere 5.2 MB. The enhanced algorithm's advantage is further demonstrated when compared to other commonly used algorithms.

Surface pressure reconstruction from LPT data with boundary conforming grids

Abstract The coupling between fluid-structure interactions is governed by the pressure distribution over the interaction surface between the fluid and solid domains. The capabilities of non-intrusive optical techniques, such as particle image velocimetry and Lagrangian particle tracking (LPT), have been proven to provide accurate velocity and acceleration information within the flow field while simultaneously tracking the corresponding structural deformations. However, scattered data from LPT measurements are typically mapped onto Cartesian grids, independently of the shape of the solid objects in the measurement domain. The use of Cartesian grids poses challenges for the determination of the surface pressure because the velocity gradients close to the object’s surface are not captured accurately. Therefore, an alternative surface pressure reconstruction scheme utilizing LPT data based on the arbitrary Lagrangian–Eulerian approach is proposed to mitigate the error propagation associated with the use of uniform grids. The introduced method provides an exact surface conformation utilizing boundary fitted coordinate systems and radial basis function based mesh deformations, which eliminates the need to use extrapolations to obtain surface pressure distributions. The introduced approach is assessed by means of a synthetic hill surface probing a three-dimensional analytical flow field; its practical applicability is demonstrated through an experimental characterization of turbulent boundary layer interactions with a steadily and unsteadily deforming elastic membrane.

Multi-receptor skin with highly sensitive tele-perception somatosensory.

The limitations and complexity of traditional noncontact sensors in terms of sensitivity and threshold settings pose great challenges to extend the traditional five human senses. Here, we propose tele-perception to enhance human perception and cognition beyond these conventional noncontact sensors. Our bionic multi-receptor skin employs structured doping of inorganic nanoparticles to enhance the local electric field, coupled with advanced deep learning algorithms, achieving a ΔV/Δd sensitivity of 14.2, surpassing benchmarks. This enables precise remote control of surveillance systems and robotic manipulators. Our long short-term memory-based adaptive pulse identification achieves 99.56% accuracy in material identification with accelerated processing speeds. In addition, we demonstrate the feasibility of using a two-dimensional (2D) sensor matrix to integrate real object scan data into a convolutional neural network to accurately discriminate the shape and material of 3D objects. This promises transformative advances in human-computer interaction and neuromorphic computing.

Three-Dimensional Dense Reconstruction: A Review of Algorithms and Datasets.

Three-dimensional dense reconstruction involves extracting the full shape and texture details of three-dimensional objects from two-dimensional images. Although 3D reconstruction is a crucial and well-researched area, it remains an unsolved challenge in dynamic or complex environments. This work provides a comprehensive overview of classical 3D dense reconstruction techniques, including those based on geometric and optical models, as well as approaches leveraging deep learning. It also discusses the datasets used for deep learning and evaluates the performance and the strengths and limitations of deep learning methods on these datasets.

Flexible Artificial Tactility with Excellent Robustness and Temperature Tolerance Based on Organohydrogel Sensor Array for Robot Motion Detection and Object Shape Recognition.

Hydrogel-based flexible artificial tactility is equipped to intelligent robots to mimic human mechanosensory perception. However, it remains a great challenge for hydrogel sensors to maintain flexibility and sensory performances during cyclic loadings at high or low temperatures due to water loss or freezing. Here, a flexible robot tactility is developed with high robustness based on organohydrogel sensor arrays with negligent hysteresis and temperature tolerance. Conductive polyaniline chains are interpenetrated through a poly(acrylamide-co-acrylic acid) network with glycerin/water mixture with interchain electrostatic interactions and hydrogen bonds, yielding a high dissipated energy of 1.58MJ m-3, and ultralow hysteresis during 1000 cyclic loadings. Moreover, the binary solvent provides the gels with outstanding tolerance from -100 to 60°C and the organohydrogel sensors remain flexible, fatigue resistant, conductive (0.27 S m-1), highly strain sensitive (GF of 3.88) and pressure sensitive (35.8 MPa-1). The organohydrogel sensor arrays are equipped on manipulator finger dorsa and pads to simultaneously monitor the finger motions and detect the pressure distribution exerted by grasped objects. A machine learning model is used to train the system to recognize the shape of grasped objects with 100% accuracy. The flexible robot tactility based on organohydrogels is promising for novel intelligent robots.

Advancing in RGB-D Salient Object Detection: A Survey

The human visual system can rapidly focus on prominent objects in complex scenes, significantly enhancing information processing efficiency. Salient object detection (SOD) mimics this biological ability, aiming to identify and segment the most prominent regions or objects in images or videos. This reduces the amount of data needed to process while enhancing the accuracy and efficiency of information extraction. In recent years, SOD has made significant progress in many areas such as deep learning, multi-modal fusion, and attention mechanisms. Additionally, it has expanded in real-time detection, weakly supervised learning, and cross-domain applications. Depth images can provide three-dimensional structural information of a scene, aiding in a more accurate understanding of object shapes and distances. In SOD tasks, depth images enhance detection accuracy and robustness by providing additional geometric information. This additional information is particularly crucial in complex scenes and occlusion situations. This survey reviews the substantial advancements in the field of RGB-Depth SOD, with a focus on the critical roles played by attention mechanisms and cross-modal fusion methods. It summarizes the existing literature, provides a brief overview of mainstream datasets and evaluation metrics, and quantitatively compares the discussed models.

Fractal-fractional model for the receptor in a 3D printing system

A critical hurdle in three-dimensional printing is the low printing accuracy. In particular, the deformation of the printed object on the receptor makes accurate printing impossible due to insufficient solvent evaporation. We address this challenge by establishing a Murray-like receptor to control the morphology of the printed object, which can be used for accurate printing through efficient residual energy absorption and active vibration attenuation. A fractal oscillator is established and He’s frequency formulation is used to reveal the fractal response of the Murray-like receptor, and an experiment was carried out to show that the receptor system can keep the printed object in good shape without any deformation.

GFENet: a lightweight network for efficient detection of abnormal objects in transmission line corridors

Abstract Abnormal objects in transmission line corridors pose a grave hazard to the security of the power system. The intelligent edge monitoring system, driven by a lightweight model, offers a promising approach. However, due to the limitations of the lightweight networks in addressing various uncertainties of anomalous objects, effectively balancing the efficiency and accuracy of existing methods presents a significant challenge. Therefore, a lightweight network called GFENet is proposed, which is designed to effectively address missed and false detections from the fluctuation in similarity of inter-class features and operational characteristics, as well as the diversity in intra-class shapes and scales under complex conditions. Firstly, learnable and efficient channel attention is proposed. This mechanism utilizes hybrid pooling fusion and weighted learning adjustment strategy to expand the receptive field, thereby capturing the distinctive visual features of the object. Next, we introduce Feature Pyramid Network and Path Aggregation Network to facilitate multi-scale feature interactions. Then, an efficient dynamic head is proposed, which employs a keypoint offset strategy to achieve scale, spatial, and task awareness. This enhances the understanding of object structure and shape without increasing computational costs. Finally, the experimental results on self-built dataset demonstrate that the GFENet can virtually balance network lightweighting and accuracy, significantly enhancing the ability to detect foreign object intrusions in complex environments.

Constructions of innovative geodesic domes in terms of the sustainable and efficient cross-sections using

Geodesic domes are structures which deliver effective solutions, associated with cost savings, due to the lack of intermediate supports when roofing large-sized objects. The multitude of advantages of this type of construction was translated into the shaping of innovative lightweight objects of geodesic domes, constructed on a regular octahedron. In this article, the use of strut sections was applied to covers generated due to a regular octahedron. Two families of domes were compared, resulting from the use of two methods that differ from each other in their topology. Each family consist of 8 structures thus finally 16 geodesic domes were considered. The generated domes were compared in terms of the same section for all strut elements, and the optimized section for each group of struts. To find the design focused on obtaining optimal solution, a number of comparative analysis were carried out. The presented analysis is extremely important in the context of environmental impact, because it shows the steel consumption and the sense of optimization. It was indicated that by optimizing the cross-sections of the strut elements, the steel consumption, as well as their weight, decreased by approx. 10–25% for dome structures which were created using method 1 and 20–40% for domes generated using method 2. The article aims to provide suggestion on the selection of the appropriate innovative geodesic dome mesh. A mesh based on various methods of shaping the covers of geodesic domes created using a regular octahedron.

Penggunaan Prototype Media Pembelajaran Untuk Anak Usia Dini Menggunakan Teknik Scan Tiga Dimensi

This research creates a prototype of learning media for early childhood using 3D scanning techniques. The prototype is designed to introduce the variety of endemic fauna in Indonesia. The research method is Analysis, Design, Development, Implementation, and Evaluation (ADDIE). This learning media is designed with five stages: Analyzing the proximity between children and technology, designing animal icons and maps, designing a 3D scan prototype using Polycam Pro, testing the 3D Scan prototype, implementation of the 3D Scan prototype. Designing a 3D Scan using Polycam Pro is better using manual methods to produce images with clear textures. 3D Scan design using Polycam Pro produces 3D shapes of endemic animal objects that are published on the web as learning media for early childhood.

CAFNet: Context aligned fusion for depth completion

Depth completion aims at reconstructing a dense depth from sparse depth input, frequently using color images as guidance. The sparse depth map lacks sufficient contexts for reconstructing focal contexts such as the shape of objects. The RGB images contain redundant contexts including details useless for reconstruction, which reduces the efficiency of focal context extraction. The unaligned contextual information from these two modalities poses a challenge to focal context extraction and further fusion, as well as the accuracy of depth completion. To optimize the utilization of multimodal contextual information, we explore a novel framework: Context Aligned Fusion Network (CAFNet). CAFNet comprises two stages: the context-aligned stage and the full-scale stage. In the context-aligned stage, CAFNet downsamples input RGB-D pairs to the scale, at which multimodal contextual information is adequately aligned for feature extraction in two encoders and fusion in CF modules. In the full-scale stage, feature maps with fused multimodal context from the previous stage are upsampled to the original scale and subsequentially fused with full-scale depth features by the GF module utilizing a dynamic masked fusion strategy. Ultimately, accurate dense depth maps are reconstructed, leveraging the GF module’s resultant features. Experiments conducted on indoor and outdoor benchmark datasets show that the CAFNet produces results comparable to state-of-the-art methods while effectively reducing computational costs.

Deocclusion and integration of advantages for a better hand pose

Estimating hand pose in the case of hand and object interaction faces the challenge of occlusion. Traditional methods that use contact information to alleviate this problem have limited applications because accurately estimating the required shape and pose of unfamiliar objects is a difficult task. This paper address this problem with occlusion removal at image level, removing the object by the proposed self-supervision method, which reduces the labor required to collect the paired labels of occlusion and deocclusion of the hand. In addition, this paper consider occlusion as valuable information and propose an integration strategy to enrich the extracted features from the occluded hand image and deoccluded hand image. Validation experiments are performed to show the proposed model’s main contributions. The experimental results on two widely-used public datasets demonstrate that the proposed model outperforms other state-of-the-art methods.

Enhancing point cloud registration with transformer: cultural heritage protection of the Terracotta Warriors

Point cloud registration technology, by precisely aligning repair components with the original artifacts, can accurately record the geometric shape of cultural heritage objects and generate three-dimensional models, thereby providing reliable data support for the digital preservation, virtual exhibition, and restoration of cultural relics. However, traditional point cloud registration methods face challenges when dealing with cultural heritage data, including complex morphological and structural variations, sparsity and irregularity, and cross-dataset generalization. To address these challenges, this paper introduces an innovative method called Enhancing Point Cloud Registration with Transformer (EPCRT). Firstly, we utilize local geometric perception for positional encoding and combine it with a dynamic adjustment mechanism based on local density information and geometric angle encoding, enhancing the flexibility and adaptability of positional encoding to better characterize the complex local morphology and structural variations of artifacts. Additionally, we introduce a convolutional-Transformer hybrid module to facilitate interactive learning of artifact point cloud features, effectively achieving local–global feature fusion and enhancing detail capture capabilities, thus effectively handling the sparsity and irregularity of artifact point cloud data. We conduct extensive evaluations on the 3DMatch, ModelNet, KITTI, and MVP-RG datasets, and validate our method on the Terracotta Warriors cultural heritage dataset. The results demonstrate that our method has significant performance advantages in handling the complexity of morphological and structural variations, sparsity and irregularity of relic data, and cross-dataset generalization.

Adjacent Image Augmentation and Its Framework for Self-Supervised Learning in Anomaly Detection.

Anomaly detection has gained significant attention with the advancements in deep neural networks. Effective training requires both normal and anomalous data, but this often leads to a class imbalance, as anomalous data is scarce. Traditional augmentation methods struggle to maintain the correlation between anomalous patterns and their surroundings. To address this, we propose an adjacent augmentation technique that generates synthetic anomaly images, preserving object shapes while distorting contours to enhance correlation. Experimental results show that adjacent augmentation captures high-quality anomaly features, achieving superior AU-ROC and AU-PR scores compared to existing methods. Additionally, our technique produces synthetic normal images, aiding in learning detailed normal data features and reducing sensitivity to minor variations. Our framework considers all training images within a batch as positive pairs, pairing them with synthetic normal images as positive pairs and with synthetic anomaly images as negative pairs. This compensates for the lack of anomalous features and effectively distinguishes between normal and anomalous features, mitigating class imbalance. Using the ResNet50 network, our model achieved perfect AU-ROC and AU-PR scores of 100% in the bottle category of the MVTec-AD dataset. We are also investigating the relationship between anomalous pattern size and detection performance.