Bounding overwatch

About: Bounding overwatch is a research topic. Over the lifetime, 966 publications have been published within this topic receiving 15156 citations.

State bounding and L∞-gain for positive singular systems with unbounded time-variable delay

Abstract: This paper studies the problems of state bounding and L ∞ -gain analysis for positive singular systems with unbounded time-variable delays and bounded disturbance for the first time. By a novel app...

Group R-CNN for Weakly Semi-supervised Object Detection with Points

Abstract: We study the problem of weakly semi-supervised object detection with points (WSSOD-P), where the training data is combined by a small set of fully annotated images with bounding boxes and a large set of weakly-labeled images with only a single point annotated for each instance. The core of this task is to train a point-to-box regressor on well-labeled images that can be used to predict credible bounding boxes for each point annotation. We challenge the prior belief that existing CNN-based detectors are not compatible with this task. Based on the classic R-CNN architecture, we propose an effective point-to-box regressor: Group R-CNN. Group R-CNN first uses instance-level proposal grouping to generate a group of proposals for each point annotation and thus can obtain a high recall rate. To better distinguish different instances and improve precision, we propose instance-level proposal assignment to replace the vanilla assignment strategy adopted in original R-CNN methods. As naive instance-level assignment brings converging difficulty, we propose instance aware representation learning which consists of instance aware feature enhancement and instance-aware parameter generation to overcome this issue. Comprehensive experiments on the MS-COCO benchmark demonstrate the effectiveness of our method. Specifically, Group R-CNN significantly outperforms the prior method Point DETR by 3.9 mAP with 5% well-labeled images, which is the most challenging scenario. The source code can be found at https://github.com/jshilong/GroupRCNN.

ObjectBox: From Centers to Boxes for Anchor-Free Object Detection

Abstract: We present ObjectBox, a novel single-stage anchor-free and highly generalizable object detection approach. As opposed to both existing anchor-based and anchor-free detectors, which are more biased toward specific object scales in their label assignments, we use only object center locations as positive samples and treat all objects equally in different feature levels regardless of the objects’ sizes or shapes. Specifically, our label assignment strategy considers the object center locations as shape- and size-agnostic anchors in an anchor-free fashion, and allows learning to occur at all scales for every object. To support this, we define new regression targets as the distances from two corners of the center cell location to the four sides of the bounding box. Moreover, to handle scale-variant objects, we propose a tailored IoU loss to deal with boxes with different sizes. As a result, our proposed object detector does not need any dataset-dependent hyperparameters to be tuned across datasets. We evaluate our method on MS-COCO 2017 and PASCAL VOC 2012 datasets, and compare our results to state-of-the-art methods. We observe that ObjectBox performs favorably in comparison to prior works. Furthermore, we perform rigorous ablation experiments to evaluate different components of our method. Our code is available at: https://github.com/MohsenZand/ObjectBox .

Real-time 3D UAV Path Planning in Dynamic Environments with Uncertainty

Abstract: The integration of Unmanned Aerial Vehicles (UAVs) is being proposed in a spectrum of applications varying from military to civil. In these applications, UAVs are required to safely navigate in real-time in dynamic and uncertain environments. Uncertainty can be present in both the UAV itself and the environment. Through a literature study, this paper first identifies, quantifies and models different uncertainty sources using bounding shapes. Then, the UAV model, path planner parameters and four scenarios of different complexity are defined. To investigate the effect of uncertainty on path planning performance, uncertainty in obstacle position and orientation and UAV position is varied between 2% and 20% for each uncertainty source first separately and then concurrently. Results show a deterioration in path planning performance with the inclusion of both uncertainty types for all scenarios for both A* and the Rapidly-Exploring Random Tree (RRT) algorithms, especially for RRT. Faster and shorter paths with similar same success rates (>95%) result for the RRT algorithm with respect to the A* algorithm only for simple scenarios. The A* algorithm performs better than the RRT algorithm in complex scenarios.

Placement of augmented reality objects using a bounding shape

Abstract: A system may include a camera, a display, one or more memories, and one or more processors communicatively coupled to the one or more memories. The system may output a bounding shape for presentation on the display. The bounding shape may be superimposed on an image being captured by the camera and presented on the display. The bounding shape may bound an object in the image. The system may determine 3D coordinates of an intersection point associated with the bounding shape. The intersection point may be a point where a projection of the bounding shape into 3D space intersects with a horizontal plane identified in the image. The system may determine 2D coordinates for presentation of an augmented reality object on the display based on the 3D coordinates of the intersection point, and may superimpose a representation of the augmented reality object on the image based on the 2D coordinates.