31 research outputs found

    Towards Confidence-guided Shape Completion for Robotic Applications

    No full text
    Many robotic tasks involving some form of 3D visual perception greatly benefit from a complete knowledge of the working environment. However, robots often have to tackle unstructured environments and their onboard visual sensors can only provide incomplete information due to limited workspaces, clutter or object self-occlusion. In recent years, deep learning architectures for shape completion have begun taking traction as effective means of inferring a complete 3D object representation from partial visual data. Nevertheless, most of the existing state-of-The-Art approaches provide a fixed output resolution in the form of voxel grids, strictly related to the size of the neural network output stage. While this is enough for some tasks, e.g. obstacle avoidance in navigation, grasping and manipulation require finer resolutions and simply scaling up the neural network outputs is computationally expensive. In this paper, we address this limitation by proposing an object shape completion method based on an implicit 3D representation providing a confidence value for each reconstructed point. As a second contribution, we propose a gradient-based method for efficiently sampling such implicit function at an arbitrary resolution, tunable at inference time. We experimentally validate our approach by comparing reconstructed shape with ground truth, and by deploying our shape completion algorithm in a robotic grasping pipeline. In both cases, we compare results with a state-of-The-Art shape completion approach. The code is available at https://github.com/andrearosasco/hyperpcr

    Defective function of Fas in patients with type-1 (insulin-dependent) diabetes mellitus associated with other autoimmune diseases

    No full text
    Fas (CD95) triggers programmed cell death and is involved in cell-mediated cytotoxicity and in shutting off the immune response. Inherited loss-of-function mutations hitting the Fas system cause the autoimmune/lymphoproliferative syndrome (ALPS). We have recently shown that ALPS patients' families display increased frequency of common autoimmune diseases, including type 1 diabetes. This work evaluates Fas function in type 1 diabetic patients without typical ALPS. Cell death induced by anti-Fas monoclonal antibody was investigated in T-cells from 13 patients with type 1 diabetes alone and 19 patients with type 1 diabetes plus other autoimmune diseases (IDDM-P). Moreover, we analyzed 19 patients with thyroiditis alone (TYR), because most IDDM-P patients displayed thyroiditis. Frequency of resistance to Fas-induced cell death was significantly higher in patients with IDDM-P (73%) than in type 1 diabetic (23%) or TYR (16%) patients or in normal control subjects (3%). The defect was specific because resistance to methyl-prednisolone-induced cell death was not significantly increased in any group. Fas was always expressed at normal levels, and no Fas mutations were detected in four Fas-resistant IDDM-P patients. Analysis of the families of two Fas-resistant patients showing that several members were Fas-resistant suggests that the defect has a genetic component. Moreover, somatic fusion of T-cells from Fas-resistant subjects and the Fas-sensitive HUT78 cell line generates Fas-resistant hybrid cells, which suggests that the Fas resistance is due to molecules exerting a dominant-negative effect on a normal Fas system. These data suggest that Fas defects may be a genetic factor involved in the development of polyreactive type 1 diabetes
    corecore