No abstract

Locating 3D objects from a single RGB image via Perspective-n-Point (PnP) is a long-standing problem in computer vision. Driven by end-to-end deep learning, recent studies suggest interpreting PnP as a differentiable layer, allowing for partial learning of 2D-3D point correspondences by backpropagating the gradients of pose loss. Yet, learning the entire correspondences from scratch is highly challenging, particularly for ambiguous pose solutions, where the globally optimal pose is theoretically non-differentiable w.r.t. the points. In this paper, we propose the EPro-PnP, a probabilistic PnP layer for general end-to-end pose estimation, which outputs a distribution of pose with differentiable probability density on the SE(3) manifold. The 2D-3D coordinates and corresponding weights are treated as intermediate variables learned by minimizing the KL divergence between the predicted and target pose distribution. The underlying principle generalizes previous approaches, and resembles the attention mechanism. EPro-PnP can enhance existing correspondence networks, closing the gap between PnP-based method and the task-specific leaders on the LineMOD 6DoF pose estimation benchmark. Furthermore, EPro-PnP helps to explore new possibilities of network design, as we demonstrate a novel deformable correspondence network with the state-of-the-art pose accuracy on the nuScenes 3D object detection benchmark. Our code is available at https://github.com/tjiiv-cprg/EPro-PnP-v2.

No abstract

III-V ternary InGaAs nanowires have great potential for electronic and optoelectronic device applications; however, the 3D structure and chemistry at the atomic-scale inside the nanowires remain unclear, which hinders tailoring the nanowires for specific applications. Here, atom probe tomography is used in conjunction with a first-principles simulation to investigate the 3D structure and chemistry of InGaAs nanowires, and reveals i) the nanowires form a spontaneous core-shell structure with a Ga-enriched core and an In-enriched shell, due to different growth mechanisms in the axial and lateral directions; ii) the shape of the core evolves from hexagon into Reuleaux triangle and grows larger, which results from In outward and Ga inward interdiffusion occurring at the core-shell interface; and iii) the irregular hexagonal shell manifests an anisotropic growth rate on {112}A and {112}B facets. Accordingly, a model in terms of the core-shell shape and chemistry evolution is proposed, which provides fresh insights into the growth of these nanowires.

Recently, automatic visual data understanding from drone platforms becomes highly demanding. To facilitate the study, the Vision Meets Drone Object Detection in Image Challenge is held the second time in conjunction with the 17-th International Conference on Computer Vision (ICCV 2019), focuses on image object detection on drones. Results of 33 object detection algorithms are presented. For each participating detector, a short description is provided in the appendix. Our goal is to advance the state-of-the-art detection algorithms and provide a comprehensive evaluation platform for them. The evaluation protocol of the VisDrone-DET2019 Challenge and the comparison results of all the submitted detectors on the released dataset are publicly available at the website: http: //www.aiskyeye.com/. The results demonstrate that there still remains a large room for improvement for object detection algorithms on drones.