Object Size Research Articles

Variable Stiffness Fibers Enabled Universal and Programmable Re-Foldability Strategy for Modular Soft Robotics.

Origami is a rich source of inspiration for creating soft actuators with complex deformations. However, implementing the re-foldability of origami on soft actuators remains a significant challenge. Herein, a universal and programmable re-foldability strategy is reported to integrate multiple origami patterns into a single soft origami actuator, thereby enabling multimode morphing capability. This strategy can selectively activate and deactivate origami creases through variable stiffness fibers. The utilization of these fibers enables the programmability of crease pattern quantity and types within a single actuator, which expands the morphing modes and deformation ranges without increasing their physical size and chamber number. The universality of this approach is demonstrated by developing a series of re-foldable soft origami actuators. Moreover, these soft origami actuators are utilized to construct a bidirectional crawling robot and a multimode soft gripper capable of adapting to object size, grasping orientation, and placing orientation. This work represents a significant step forward in the design of multifunctional soft actuators and holds great potential for the advancement of agile and versatile soft robots.

Laboratory-based x-ray dark-field microscopy.

We demonstrate the capability of laboratory-based x-ray microscopes, using intensity-modulation masks, to access the sub-micron length scale in the dark field contrast channel while maintaining micron resolution in the resolved (refraction and attenuation) channels. The dark field contrast channel reveals the presence of ensembles of samples' features below the system resolution. Resolved refraction and attenuation channels provide multi-modal high-resolution imaging down to the micron scale. We investigate the regimes of modulated and un-modulated dark field as well as refraction, quantifying their dependence on the relationship between feature size in the imaged object and aperture size in the intensity-modulation mask. We propose an analytical model to link the measured signal with the sample's microscopic properties. Finally, we demonstrate the relevance of the microscopic dark field contrast channel in applications from both the life and physical sciences, providing proof of concept results for imaging collagen bundles in cartilage and dendritic growth in lithium batteries.

Bimodal SegNet: Fused instance segmentation using events and RGB frames

Object segmentation enhances robotic grasping by aiding object identification. Complex environments and dynamic conditions pose challenges such as occlusion, low light conditions, motion blur and object size variance. To address these challenges, we propose a Bimodal SegNet that fuses two types of visual signals, event-based data and RGB frame data. The proposed Bimodal SegNet network has two distinct encoders — one for RGB signal input and another for Event signal input, in addition to an Atrous Pyramidal Feature Amplification module. Encoders capture and fuse the rich contextual information from different resolutions via a Cross-Domain Contextual Attention layer while the decoder obtains sharp object boundaries. The evaluation of the proposed method undertakes five unique image degradation challenges including occlusion, blur, brightness, trajectory and scale variance on the Event-based Segmentation (ESD) Dataset. The results show a 4%–6% MIOU score improvement over state-of-the-art methods in terms of mean intersection over the union and pixel accuracy. The source code, dataset and model are publicly available at: https://github.com/sanket0707/Bimodal-SegNet.

An Efficient Rep-Style Gaussian–Wasserstein Network: Improved UAV Infrared Small Object Detection for Urban Road Surveillance and Safety

Owing to the significant application potential of unmanned aerial vehicles (UAVs) and infrared imaging technologies, researchers from different fields have conducted numerous experiments on aerial infrared image processing. To continuously detect small road objects 24 h/day, this study proposes an efficient Rep-style Gaussian–Wasserstein network (ERGW-net) for small road object detection in infrared aerial images. This method aims to resolve problems of small object size, low contrast, few object features, and occlusions. The ERGW-net adopts the advantages of ResNet, Inception net, and YOLOv8 networks to improve object detection efficiency and accuracy by improving the structure of the backbone, neck, and loss function. The ERGW-net was tested on a DroneVehicle dataset with a large sample size and the HIT-UAV dataset with a relatively small sample size. The results show that the detection accuracy of different road targets (e.g., pedestrians, cars, buses, and trucks) is greater than 80%, which is higher than the existing methods.

Ghost Imaging Through a Supersonic Wind‐Induced Environment Under Weak Illumination

AbstractIn airborne imaging, wind‐induced random environment severely disturbs the imaging quality. Ghost imaging (GI) shows its potential application in airborne imaging due to the natural ability of anti‐interference and imaging at low light levels, however, the influence of wind‐induced random environment on GI is non‐negligible under low‐light detection. In this paper, GI through a wind‐induced random environment under low‐light detection is experimentally and numerically investigated. The experiment is performed with a wind tunnel, and the imaging quality decreases with wind velocity and detection signal‐to‐noise ratio (SNR). The simulation model of GI under low‐light detection through this kind of environment is proposed, and simulation results match well with experiments. With the theoretical model, imaging results for the binary and grayscale object under supersonic wind region are discussed in detail, and the critical wind velocity under the influences of boundary‐layer thickness and object size are obtained. These results can find potential applications in optical imaging and can be used to estimate the effect of airflow on the imaging performance.

Fusion of dual modalities of non-invasive sensory feedback for object profiling with prosthetic hands.

Either non-invasive somatotopic or substitute sensory feedback is capable of conveying a single modality of sensory information from prosthetic hands to amputees. However, the neurocognitive ability of amputees to integrate multi-modality sensory information for functional discrimination is unclear. The purpose of this study was to assess the fusion of non-invasive somatotopic tactile and substitute aperture feedbacks for profile perception of multiple physical features during grasping objects. Two left transradial amputees with somatotopic evoked tactile sensation (ETS) of five fingers participated in the study. The tactile information of prosthetic hand was provided to amputees by the ETS feedback elicited on the stump projected finger map. Hand aperture information was conveyed to amputees with substitute electrotactile stimulation on the forearm or upper arm. Two types of sensory feedback were integrated to a commercial prosthetic hand. The efficacy of somatotopic ETS feedback on object length identification task was assessed with or without substitute aperture stimulation. The object size identification task was utilized to assess how ETS stimulation at the stump may affect aperture perception with stimulation on the ipsilateral upper arm or forearm. Finally, the task of identifying combined length and size was conducted to evaluate the ability of amputees to integrate the dual modalities of sensory feedback for perceiving profile features. The study revealed that amputee subjects can effectively integrate the ETS feedback with electrotactile substitutive feedback for object profile discrimination. Specifically, ETS was robust to provide object length information with electrotactile stimulation at either the forearm or upper arm. However, electrotactile stimulation at the upper arm for aperture perception was less susceptible to the interference of ETS stimulation than at the forearm. Amputee subjects are able to combine somatotopic ETS and aperture feedbacks for identifying multi-dimensional features in object profiling. The two sensory streams of information can be fused effectively without mutual interference for functional discrimination.

Deep Learning-Based Consistent Object Detection in Distance on Highly Perspective Region

CCTVs are commonly used for traffic monitoring and accident detection, but their images suffer from severe perspective distortion causing object size reduction with distance. This issue is exacerbated in tunnel CCTVs, positioned low due to space constraints, leading to challenging object detection, especially for distant small objects, due to perspective effects. To address this, this study proposes a solution involving a region of interest setup and an inverse perspective transformation technique. The transformed images, achieved through this technique, enlarge distant objects, maintaining object detection performance and appearance velocity across distances. To validate this, artificial CCTV images were generated in a virtual tunnel environment, creating original and transformed image datasets under identical conditions. Comparisons were made between the appearance velocity and object size of individual vehicles and for deep learning model performance with multiple moving vehicles. The evaluation was conducted across four distance intervals (50 m to 200 m) from the tunnel CCTV location. The results reveal that the model using original images experiences a significant decline in object detection performance beyond 100 m, while the transformed image-based model maintains a consistent performance up to the distance of 200 m.

Anchor boxes adaptive optimization algorithm for maritime object detection in video surveillance

With the development of the marine economy, video surveillance has become an important technical guarantee in the fields of marine engineering, marine public safety, marine supervision, and maritime traffic safety. In video surveillance, maritime object detection (MOD) is one of the most important core technologies. Affected by the size of maritime objects, distance, day and night weather, and changes in sea conditions, MOD faces challenges such as false detection, missed detection, slow detection speed, and low accuracy. However, the existing object detection algorithms usually adopt predefined anchor boxes to search and locate for objects of interest, making it difficult to adapt to maritime objects’ complex features, including the varying scale and large aspect ratio difference. Therefore, this paper proposes a maritime object detection algorithm based on the improved convolutional neural network (CNN). Firstly, a differential-evolutionary-based K-means (DK-means) anchor box clustering algorithm is proposed to obtain adaptive anchor boxes to satisfy the maritime object characteristics. Secondly, an adaptive spatial feature fusion (ASFF) module is added in the neck network to enhance multi-scale feature fusion. Finally, focal loss and efficient intersection over union (IoU) loss are adopted to replace the original loss function to improve the network convergence speed. The experimental results on the Singapore maritime dataset show that our proposed algorithm improves the average precision by 7.1%, achieving 72.7%, with a detection speed of 113 frames per second, compared with You Only Look Once v5 small (YOLOv5s). Moreover, compared to other counterparts, it can achieve a better speed–accuracy balance, which is superior and feasible for the complex maritime environment.

Farmed Atlantic salmon (Salmo salar L.) avoid intrusive objects in cages: The influence of object shape, size and colour, and fish length

Underwater vehicles and other mobile platforms are seeing increased use as tools within fish farming, particularly due to current trends towards Precision Farming practices, and more exposed farming sites. Although many of the applications of such tools (e.g., net cleaning and inspection) have become well established industrial practices, it is largely unknown how much such operations disturb the fish and the consequence of this disturbance. In this study, we explored this by exposing Atlantic salmon in commercial net cages to intrusive objects and monitoring the distribution of fish around these using on-board 360-degree sonars. Six different object designs were tested covering variations in size, shape, and colour, which are important static characteristics of underwater vehicles/platforms. The sonar data was first aggregated into images containing the Cumulative Fish Presence over 1-, 5- and 10-min periods to provide a more robust foundation for further analyses. By training a deep learning based method using UNet++ architecture to automatically segment the fish distribution patterns, the mean distance between the inner perimeter of the fish distribution and the object was assessed. Results from the study implied that fish keep greater distances to larger objects. There was, however, no clear impact of the shape. Regarding the effect of colour, fish kept greater distances to yellow than to white objects. When comparing results from tests on fish of different size, data indicate a positive linear relationship between fish weight (age) and distance to an object, that can be expressed as an avoidance distance of an average 3.8 body-lengths. Our findings provide new fundamental knowledge on the dynamics between the fish and objects such as vehicles or other mobile platforms in fish farms, and thus provides valuable insights that can be useful when designing such tools specifically for aquaculture.

Indoor semantic segmentation based on Swin-Transformer

In recent years, with the rapid development of Transformer in the field of natural language processing, many researchers have realized its potential and gradually applied it to the field of computer vision, with a proliferation of theoretical approaches represented by Vision Transformer (ViT) and Data-efficient image Transformer (DeiT). On the basis of ViT, the famous Swin-Transformer was proposed as one of the best computer vision neural network backbones, which can be widely used in tasks such as image classification, target detection and video recognition. However, in the field of image segmentation, the semantic segmentation of indoor scenes is still very challenging due to the wide variety of objects, large differences in object sizes, and a large number of overlapping objects with occlusion. Aiming at the problem that the existing semantic segmentation of RGB-D indoor scenes cannot effectively fuse multimodal features, in this paper, we propose a novel indoor semantic segmentation algorithm based on Swin-Transformer. It attempts to apply Swin-Transformer to the field of indoor RGBD semantic segmentation, and tests the performance of the model by conducting extensive experiments on the mainstream indoor semantic segmentation datasets NYU-Depth V2 and SUN RGB-D. The experimental results show that the Swin-L RGB+Depth setting achieves 52.44% MIoU on the NYU-Depth V2 data and 51.15% MIoU on the SUN RGB-D data set, which reflects an excellent performance in the field of indoor semantic segmentation. The improved performance of the Depth features on the indoor semantic segmentation model has also been demonstrated in the experiments by controlling the type of input features. Our source code is publicly available at https://github.com/YunpingZheng/ISSSW.

Evaluation of Changes in Dose Estimation on Abdomen CT Scan with Automatic Tube Current Modulation Using In-House Phantom

This study evaluates the effect of the Automatic Tube Current Modulation (ATCM) technique on pitch and effective diameter variation in estimating dose values and noise levels for abdominal examination on Philips Ingenuity CT scan machine using in-house Phantoms. The in-house phantoms are oval in shape with three effective diameter sizes, namely 23.2 cm, 28.3 cm, and 33.3 cm to represent abdominal region. The three size Phantoms were scanned using an Ingenuity 128 Philips CT scan with the abdominal protocol exposure parameters of 120 kVp tube voltage, Dose Right Index (DRI) variations of 10,11,12,13, and 14, and pitch variations of 0.6; 0.8; 1.0; 1.2; and 1.49. The changes in mAs, CTDIvol, and noise to the Philips reference value were then verified (i.e. an addition of one DRI value increases mAs by 12 %). For evaluation, a metric to express the change in DRI is defined as ΔDRI. The study demonstrates that noise level is influenced by object size; size information of the object could be useful to predict the change of tube current and pitch due to ATCM with respect to selected DRI. The DRI value is proportional to the tube current, thus selecting the DRI at a certain pitch will directly determine tube current. The ΔDRI in general, according to Philips specifications, is verified to be approximately 10 % to 13 %, except for DRI 10 to 11 which is relatively high on average 15 % to 17 %. Increasing DRI increases the CTDIvol. The CTDI/mAs constantly ranges of 0.06 to 0.07. The value could serve as a characteristic parameter for quality assurance. The ATCM specifications of the Ingenuity 128 CT Scanner is according to Philips regulations.

A Historical Review of Research on the Density-Size Correlation of the Observable Universe

This literature review explores the correlation between the density and size of celestial objects - a key subject in understanding their formation, evolution, and the potential existence of higher-order clusters. It draws attention to the two seminal papers - E. F. Carpenter's density-size correlation study and Gerard De Vaucouleurs's study on matter density in groups and clusters. While acknowledging the limitations and outdated nature of Carpenter's body of work, it underscores the need for a more contemporary, comprehensive examination across various scales from neutron stars to galaxy clusters. The review highlights how such a study can clarify questions surrounding the size and shape of the universe, address Olbers' Paradox, and contribute to theories that describe the universe's origins and dynamics.

Spatial resolution and reconstructed size accuracy using advanced beamformers in linear array-based PAT systems

Limitations associated with linear-array probes in photoacoustic tomography are partially compensated by using advanced beamformers that exploit the temporal and spatial coherence of the recorded signals, such as Delay Multiply and Sum (DMAS), Minimum Variance (MV) or coherence factor (CF), among others. However, their associated signal processing leads to an overestimation of the spatial resolution, as well as alterations in the reconstructed object size. Numerical and experimental results reported here support this hypothesis. First, we show that the Rayleigh criterion (RC) is the most suitable choice to characterize the spatial resolution instead of the Point Spread Function (PSF) when considering advanced beamformers. Then, we observe that several advanced beamformers fail to properly reconstruct target sizes slightly above the spatial resolution, underestimating their size. This work sheds light on the suitability of this type of beamformers combined with linear probes for determining sizes and morphology in photoacoustic images.

Towards Robust Obstacle Avoidance for the Visually Impaired Person Using Stereo Cameras

We propose a novel obstacle avoidance strategy implemented in a wearable assistive device, which serves as an electronic travel aid (ETA), designed to enhance the safety of visually impaired persons (VIPs) during navigation to their desired destinations. This method is grounded in the assumption that objects in close proximity and within a short distance from VIPs pose potential obstacles and hazards. Furthermore, objects that are farther away appear smaller in the camera’s field of view. To adapt this method for accurate obstacle selection, we employ an adaptable grid generated based on the apparent size of objects. These objects are detected using a custom lightweight YOLOv5 model. The grid helps select and prioritize the most immediate and dangerous obstacle within the user’s proximity. We also incorporate an audio feedback mechanism with an innovative neural perception system to alert the user. Experimental results demonstrate that our proposed system can detect obstacles within a range of 20 m and effectively prioritize obstacles within 2 m of the user. The system achieves an accuracy rate of 95% for both obstacle detection and prioritization of critical obstacles. Moreover, the ETA device provides real-time alerts, with a response time of just 5 s, preventing collisions with nearby objects.

Deer survey from drone thermal imagery using enhanced faster R-CNN based on ResNets and FPN

Deer surveys play an important role in the estimation of local ecological balance. In the Chitwan National Park of Nepal, the dense tree canopies and tall vegetation often obscure the presence of wild deer, which has a negative effect on the accurate population surveys of wild deer. UAVs equipped with infrared sensors have been increasingly used to monitor wild deer by capturing a lot of images. How to automatically recognize and obtain the number of deer objects from thermal images is becoming an important research topic. Due to the difference between thermal images and true-color images, as well as the variations in deer object sizes in these two types of images, current ready-to-use object detection models, designed for true-color imagery, are ill-suited for the task of detecting small deer objects within thermal imagery. In this paper, an enhanced Faster R-CNN was constructed to detect small deer objects from thermal images, in which a Feature Pyramid Network (FPN) based on a residual network is used to improve feature extraction for small deer objects and multi-scale feature map constrution for the subsequent region proposals searching, bounding box regression, and regions of interest (RoIs) classification. In addition, small-scaled anchor boxes and a multi-scale feature map selection criterion are devised to improve the detection accuracy of small objects. Finally, based on Faster R-CNN, FPN, and different residual networks including ResNet18, ResNet34, ResNet50, ResNet101, and ResNet152, we constructed five object detection models, and evaluated their detection performance by using COCO evaluation matrix. Under the condition of IoU≥0.5, the integration of Faster R-CNN, FPN, and ResNet18 demonstrated to perform better than others. Specifically, The COCO evaluation results revealed an Average Precision (AP) score of 91.6% for all deer objects. Small deer objects (area ≤ 200 pixels) achieved an AP score of 73.6%, medium deer objects (200 < area ≤ 400 pixels) demonstrated an AP score of 93.4%, and large deer objects (area > 400 pixels) achieved the highest AP score of 94.3%. Our research is helpful for effective wild deer monitoring and conservation and can be a valuable reference for the exploration of small object detection from low-resolution thermal images.

Reduced perception-action dissociation in children with amblyopia

The functional distinction between vision-for-perception and vision-for-action is a key aspect of understanding the primate visual system. While this dissociation has been well-established in adulthood, its development and dependence on typical visual experience remain unclear. To address these questions, we examined two groups of children: typically developed children and those with amblyopia, who presumably have a sub-optimal visual experience. The Ponzo illusion, known to impact perception but not visuomotor behaviors across age groups, was employed to assess the extent of dissociation. Participants engaged in two tasks involving the Ponzo illusion: a grasping task (vision-for-action) and a manual estimation task (vision-for-perception), with objects placed on the "close" and "far" surfaces of the illusion. Typically developed children displayed grasping movements that were unaffected by the illusion, as their grasping apertures were scaled based on object size, independent of its location.In contrast, children with amblyopia exhibited a clear susceptibility to the illusion, showing larger apertures for objects placed on the 'far' surface of the illusion, and smaller apertures for objects placed on the 'close' surface. Interestingly, both groups of children demonstrated similar susceptibility to the illusion during the perceptual task, with objects placed on the far surface being perceived as longer compared to objects placed on the close surface. These findings shed light on the impact of atypical visual development on the emergence of the dissociation between perception and action, highlighting the crucial role of typical visual experience in establishing this distinction.

Crystal plasticity simulations with representative volume element of as-build laser powder bed fusion materials

Additive manufacturing of as-build metal materials with laser powder bed fusion typically leads to the formations of various chemical phases and their corresponding microstructure types. Such microstructures have very complex shape and size anisotropic distributions due to the history of the laser heat gradients and scanning patterns. With higher complexity compared to the post-heat-treated materials, the synthetic volume reconstruction of as-build materials for accurate modelling of their mechanical properties is a serious challenge. Here, we present an example of complete workflow pipeline for such nontrivial task. It takes into account the statistical distributions of microstructures: object sizes for each phase, several shape parameters for each microstructure type, and their morphological and crystallographic orientations. In principle, each step in the pipeline, including the parameters in the crystal plasticity model, can be fine-tuned to achieve suitable correspondence between experimental and synthetic microstructures as well as between experimental stress–strain curves and simulated results. To our best knowledge, this work represents an example of the most challenging synthetic volume reconstruction for as-build additive manufacturing materials to date.

Drawing as a tool for investigating the nature of imagery representations of blind people: The case of the canonical size phenomenon.

Several studies have shown that blind people, including those with congenital blindness, can use raised-line drawings, both for "reading" tactile graphics and for drawing unassisted. However, research on drawings produced by blind people has mainly been qualitative. The current experimental study was designed to investigate the under-researched issue of the size of drawings created by people with blindness. Participants (N = 59) varied in their visual status. Adventitiously blind people had previous visual experience and might use visual representations (e.g., when visualising objects in imagery/working memory). Congenitally blind people did not have any visual experience. The participant's task was to draw from memory common objects that vary in size in the real world. The findings revealed that both groups of participants produced larger drawings of objects that have larger actual sizes. This means that the size of familiar objects is a property of blind people's mental representations, regardless of their visual status. Our research also sheds light on the nature of the phenomenon of canonical size. Since we have found the canonical size effect in a group of people who are blind from birth, the assumption of the visual nature of this phenomenon - caused by the ocular-centric biases present in studies on drawing performance - should be revised.

Visual imagery deficits in posterior cortical atrophy

ABSTRACT Visual imagery has a close overlapping relationship with visual perception. Posterior cortical atrophy (PCA) is a neurodegenerative syndrome marked by early impairments in visuospatial processing and visual object recognition. We asked whether PCA would therefore also be marked by deficits in visual imagery, tested using objective forced-choice questionnaires, and whether imagery deficits would be selective for certain properties. We recruited four patients with PCA and a patient with integrative visual agnosia due to bilateral occipitotemporal strokes for comparison. We administered a test battery probing imagery for object shape, size, colour lightness, hue, upper-case letters, lower-case letters, word shape, letter construction, and faces. All subjects showed significant impairments in visual imagery, with imagery for lower-case letters most likely to be spared. We conclude that PCA subjects can show severe deficits in visual imagery. Further work is needed to establish how frequently this occurs and how early it can be found.

Simulated eye height impacts size perception differently depending on real-world posture

Changes in perceived eye height influence visually perceived object size in both the real world and in virtual reality. In virtual reality, conflicts can arise between the eye height in the real world and the eye height simulated in a VR application. We hypothesized that participants would be influenced more by variation in simulated eye height when they had a clear expectation about their eye height in the real world such as when sitting or standing, and less so when they did not have a clear estimate of the distance between their eyes and the real-life ground plane, e.g., when lying supine. Using virtual reality, 40 participants compared the height of a red square simulated at three different distances (6, 12, and 18 m) against the length of a physical stick (38.1 cm) held in their hands. They completed this task in all combinations of four real-life postures (supine, sitting, standing, standing on a table) and three simulated eye heights that corresponded to each participant’s real-world eye height (123cm sitting; 161cm standing; 201cm on table; on average). Confirming previous results, the square’s perceived size varied inversely with simulated eye height. Variations in simulated eye height affected participants’ perception of size significantly more when sitting than in the other postures (supine, standing, standing on a table). This shows that real-life posture can influence the perception of size in VR. However, since simulated eye height did not affect size estimates less in the lying supine than in the standing position, our hypothesis that humans would be more influenced by variations in eye height when they had a reliable estimate of the distance between their eyes and the ground plane in the real world was not fully confirmed.