To satisfy the rigorous requirements of precise edge detection in critical high-accuracy measurements, this article proposes a series of efficient approaches for localizing subpixel edge. In contrast to the fitting ba...
详细信息
Active vision enables dynamic and robust visual perception, offering an alternative to the static, passive nature of feedforward architectures commonly used in computer vision, which depend on large datasets and high ...
详细信息
Active vision enables dynamic and robust visual perception, offering an alternative to the static, passive nature of feedforward architectures commonly used in computer vision, which depend on large datasets and high computational resources. Biological selective attention mechanisms allow agents to focus on salient Regions of Interest (ROIs), reducing computational demand while maintaining real-time responsiveness. Event-based cameras, inspired by the mammalian retina, further enhance this capability by capturing asynchronous scene changes, enabling efficient, low-latency processing. To distinguish moving objects while the event-based camera is also in motion, the agent requires an object motion segmentation mechanism to accurately detect targets and position them at the center of the visual field (fovea). Integrating event-based sensors with neuromorphic algorithms represents a paradigm shift, using Spiking Neural Networks (SNNs) to parallelise computation and adapt to dynamic environments. This work presents a Spiking Convolutional Neural Network (sCNN) bioinspired attention system for selective attention through object motion sensitivity. The system generates events via fixational eye movements using a Dynamic Vision Sensor (DVS) integrated into the Speck neuromorphic hardware, mounted on a Pan-Tilt unit, to identify the ROI and saccade toward it. The system, characterised using ideal gratings and benchmarked against the Event Camera Motion Segmentation Dataset (EVIMO), reaches a mean IoU of 82.2% and a mean SSIM of 96% in multi-object motion segmentation. Additionally, the detection of salient objects reaches an accuracy of 88.8% in office scenarios and 89.8% in challenging indoor and outdoor low-light conditions, as evaluated on the Event-Assisted Low-Light Video Object Segmentation Dataset (LLE-VOS). A real-time demonstrator showcases the system’s capabilities of detecting the salient object through object motion sensitivity in 0.124 s in dynamic scenes. Its l
Over the past decade, cubic boron arsenide (BAs) has emerged as a highly promising semiconductor owing to its extraordinary thermal conductivity (1,200 W/m·K) and high ambipolar mobility (1,600 cm2/V·s). Thi...
详细信息
Evaluating trunk control ability is significant in guiding patients towards proper functional training. Existing assessment techniques are subjective with low resolution, lack multi-dimensional assessment capability, ...
详细信息
Point cloud data frames are critical, if not indispensable, for precise robot navigation and localization, but training the object detection models for them remains challenging. Many models require labeled 3D objects ...
详细信息
ISBN:
(纸本)9798400706295
Point cloud data frames are critical, if not indispensable, for precise robot navigation and localization, but training the object detection models for them remains challenging. Many models require labeled 3D objects to train the model. However, the sparse and occluded 3D point cloud data make it difficult, if not impossible, to automate the labeling process. This work proposes a training framework to generate 3D labels on point clouds to tackle the aforementioned challenges. The proposed method takes advantage of the consecutive presence of the same object on different frames to automate the labeling process. The experimental results show that the unsupervised framework trains a robust model for 3D object detection. On the roadside data, the model archives 90.27% AP for scooters and 91.33% AP for cars. On nuScenes dataset, the framework demonstrates the detection precision doubles on IoU 0.25 and IoU 0.5 when the recalls remain similar, compared to the model trained by the MODEST framework.
This Letter presents a combined analytical and experimental method to effectively decouple the radial and tangential residual stress fields induced by Berkovich nanoindentation in single-crystalline 4H-SiC using micro...
详细信息
This paper presents the kinematic and dynamic analysis of an n -link manipulator with flexible members. The deformation of a link from its rigid body position is modeled by a homogeneous transformation. The dynamic eq...
详细信息
This paper presents the kinematic and dynamic analysis of an n -link manipulator with flexible members. The deformation of a link from its rigid body position is modeled by a homogeneous transformation. The dynamic equations are obtained using Euler-Lagrange formulation. These equations are compared to those describing a rigid link mechanism.
The main objective is to provide an evidence of spatial dependence of mechanical responses of a heterogeneous aluminum brazed joint re-solidified clad, and to confirm a sufficient sensitivity of a nano-indentation—lo...
详细信息
The main objective is to provide an evidence of spatial dependence of mechanical responses of a heterogeneous aluminum brazed joint re-solidified clad, and to confirm a sufficient sensitivity of a nano-indentation—load curve method for identifying the dependence. Topological features of a network of solidification microstructures(α phase and eutectic), formed during quench in a brazing process of aluminum alloy, influence significantly dynamic mechanical responses of resulting heterogeneous material. Nano/micro indentation depth vs load characteristics of differing phases suggest a spatially sensitive mechanical response of a re-solidified fillet in the joint zone. Hence, a spatial distribution, pattern formations and other morphological characteristics of microstructures have a direct impact on an ultimate joint integrity. Topology-induced variations of indentation—load curves was presented. A hypothesis involving microstructures’ spatial distribution vs mechanical response was formulated.
A dynamic model that represents an exact linearization scheme with a simplified nonlinear feedback is presented. To realize this model for robotic systems, the output functions should be chosen so that a special decom...
详细信息
A dynamic model that represents an exact linearization scheme with a simplified nonlinear feedback is presented. To realize this model for robotic systems, the output functions should be chosen so that a special decomposition of the total inertial matrix is satisfied. The concept of an imaginary robot is utilized to achieve the formulation and to solve the realization problem. Two illustrative examples are given in the paper, one for the Stanford arm and the other for a PUMA type of robot. An optimal robotic physical design and a control system design based on the new model are also discussed.
暂无评论