Efficient 3D object recognition using multifoveated point cloud
Foveated structure, multifoveation, 3D object recognition, point cloud, foveated point cloud
Technological innovations in the field of hardware and RGB-D sensors allowed to realize the acquisition of 3D point clouds in real time. As a consequence, have been arisen varieties of interactive applications related to the 3D world that have been receiving increasing attention from researchers. However, one of the main problems that still remains is the demand for computationally intensive processing that requires optimized approaches to deal with this 3D vision model, especially when it is necessary to perform tasks in real time. However, one of the main problems that still remains is the demand for computationally intensive processing that requires optimized approaches to deal with this 3D vision model, especially when it is necessary to perform tasks in real time. Thus, we start from the proposed 3D multiresolution model presented as foveated point clouds which is a possible solution to this problem, but is limited to a single foveated structure with context dependent mobility. In this way, our proposal is an improvement of this model with the incorporation of multiple foveated structures. In the context of multifoveation, the application of several foveated structures results in a considerable increase of processing, since there are intersections between regions of distinct structures, which are processed multiple times. The proposal also brings an approach that avoids the processing of redundant regions, resulting in reduced processing time. Such an approach can be used to identify objects in 3D point clouds, one of the key tasks for automation, with efficient synchronization, allowing the validation of the model and verification of its applicability in the context of computer vision. The partial results demonstrate a gain in performance of the proposed model in relation to the use of multiple foveated point cloud structures.