Brain tumors pose a significant health risk to humans. For proper diagnosis and efficient treatment planning, early detection of brain tumors is crucial. Nevertheless, current automated detection techniques face several challenges, including insufficient robustness in complex environments, difficulties in recognizing boundaries, and limitations in identifying small tumors. To solve these problems, this paper proposes a novel YOLO-BT architecture that integrates the inverted Residual Mobile Block (iRMB), the Spatial and Channel Reconstruction Convolution (ScConv) module, and the Wise-IoU loss function into YOLOv9. Our experimental results demonstrate that on the Br35H dataset, YOLO-BT achieved a 1.6% improvement in [email protected] compared to YOLOv9, while on the Axial-T1CE-2-Class dataset, YOLO-BT exhibited a 2.2% improvement over YOLOv9. This demonstrates that YOLO-BT is an effective and practical solution for brain tumor detection.

Perspective-taking and attentional switching are some of the ergonomic challenges that existing teleoperation human-machine interface designs need to address. This study developed two gaze interaction methods, the Eye Stick and the Eye Click, which were based on the joystick metaphor and the navigation metaphor, respectively, to be used in exocentric perspective teleoperation scenarios. We conducted two user studies to test the task performance and the subjective experience of the gaze interaction methods in a virtual ground vehicle teleoperation task. The results showed that compared with a traditional joystick design, the Eye Stick led to a shorter driving distance and the Eye Click led to less task time, and the gaze interaction methods had performance advantages in more difficult mazes. After multiple task sessions, the participants reported that the gaze interaction methods and the traditional joystick were similar in terms of task workload, perceived learnability, and satisfaction; however, the perceived usability of the Eye Stick was not as good as the Eye Click and the traditional joystick. In conclusion, both the Eye Stick and the Eye Click are feasible and promising gaze interaction methods for teleoperation applications with task performance advantages; however, more research is needed to optimize their user experience design.

Previous studies have used the Fitts’ Law to predict the performance of pointing task on touch screens. It was found that moving distance, target width, and the direction of motion would affect the task performance. The present research aimed to investigate the impact of the initial and terminal positions (ITPs) on pointing performance in two Exps. ITPs were divided into two categories: center of a screen and outer of a screen (center ↔ outer) pointing, and outer of a screen to outer of a screen (outer→outer) pointing. In Exp 1, 30 participants performed the center ↔ outer pointing tasks with 8 directions. The results showed that the outer → center movement was significantly faster than the center → outer at 45°, 90°, and 180°. In Exp 2, 30 participants performed the outer → outer pointing tasks with eight directions. The current study revealed that the ITPs influenced the performance of the pointing task, possibly due to human biomechanical characteristics associated with different movements.

Haptic rendering is a highly device-dependent process. Accordingly, haptic rendering algorithms should be designed regarding the specific physical features of a haptic device. The Immersion CyberGrasp is a force feedback device which can provide an individual force roughly perpendicular to the fingertip of each finger. In this paper, we propose a novel force rendering algorithm for the Immersion CyberGrasp considering the feature of area contacts between the virtual hand avatar and target geometric models. Line segment clusters are used as haptic proxies for the real-time collision detection between the fingertips and target geometric models. Force computations are conducted directly using the collision detection results. Spatial subdivision is performed on target geometric models to facilitate real-time collision detection. Experimental results show that our algorithm can provide stable and realistic force feedback.

Intuitive and natural interactions are key aspects of virtual assembly applications. In this paper, we present intuitive virtual hand interactions with parts, subassemblies and tools. Grasping patterns are developed based on the abstraction of various mechanic shapes, kinematics model of virtual hand is built upon the anatomic structure of the human hand to guarantee motion realism, and real-time collision detection scheme is used to identify whether the current finger configuration matches a specific grasping pattern. Experiments on assembly models with a CyberGlove dataglove show that our virtual grasping scheme works well for virtual assembly tasks.

In this paper we present an interactive collision detection algorithm for complex and deformable objects. For two target models, our approach rapidly calculates their region of interests (ROI), which is the overlapping of their axis aligned bounding boxes (AABBs), in CPU. The surfaces of both models inside the ROI are then voxelized using a novel GPU-based real-time voxelization method. The resultant volumes are represented by two 2D textures in video memory. The collision query is efficiently accomplished by comparing these 2D textures in GPU. The algorithm is robust to handle arbitrary shapes, no matter geometric models are convex or concave, closed or open, rigid or deformable. Our preliminary implementation achieves interactive frame rate for complex models with up to one million triangles on commodity desktop PCs.

Abstract Image‐based collision detection algorithms make efficient use of the graphics rendering hardware and reduce the computational cost of CPU. In this article, a fast collision detection algorithm based on image space is presented, which combines graphics hardware capabilities with a simplified geometric representation (oriented bounding box) in order to rapidly detect collisions between complex objects. The method can deal with arbitrary polyhedra, while preserving the merits of image‐based collision detection algorithms. This is achieved by decomposing the surfaces of an object into a list of convex pieces. High efficiency of the algorithm is obtained by organizing the convex pieces into a binary tree with each node representing a convex piece, and by adopting triangle strip compression. The algorithm has been implemented and compared with related algorithms. Copyright © 2003 John Wiley & Sons, Ltd.