Blended Convolution and Synthesis for Efficient Discrimination of 3D Shapes
This addresses the challenge of analyzing irregular 3D point clouds for shape analysis, offering a lightweight solution that is incremental in improving efficiency and discrimination.
The paper tackles the problem of inefficient inter-class discrimination in 3D point clouds by proposing a 'Blended Convolution and Synthesis' layer that synthesizes compact representations and uses a novel 3D convolution operator, achieving compelling results on 3D shape recognition and retrieval.
Existing networks directly learn feature representations on 3D point clouds for shape analysis. We argue that 3D point clouds are highly redundant and hold irregular (permutation-invariant) structure, which makes it difficult to achieve inter-class discrimination efficiently. In this paper, we propose a two-faceted solution to this problem that is seamlessly integrated in a single `Blended Convolution and Synthesis' layer. This fully differentiable layer performs two critical tasks in succession. In the first step, it projects the input 3D point clouds into a latent 3D space to synthesize a highly compact and more inter-class discriminative point cloud representation. Since, 3D point clouds do not follow a Euclidean topology, standard 2/3D Convolutional Neural Networks offer limited representation capability. Therefore, in the second step, it uses a novel 3D convolution operator functioning inside the unit ball ($\mathbb{B}^3$) to extract useful volumetric features. We extensively derive formulae to achieve both translation and rotation of our novel convolution kernels. Finally, using the proposed techniques we present an extremely light-weight, end-to-end architecture that achieves compelling results on 3D shape recognition and retrieval.