Intrinsic Image Decomposition Using Point Cloud Representation

TL;DR

This paper introduces PoInt-Net, a novel 3D point cloud-based method for intrinsic image decomposition that outperforms traditional 2D approaches in accuracy, efficiency, and robustness across diverse scenes.

Contribution

The paper presents a new point cloud-based neural network for intrinsic decomposition, capable of generalizing well and operating efficiently on various scene sizes.

Findings

Achieves superior accuracy over 2D methods.

Demonstrates robustness to unseen objects and scenes.

Efficiently handles point clouds of any size.

Abstract

The purpose of intrinsic decomposition is to separate an image into its albedo (reflective properties) and shading components (illumination properties). This is challenging because it's an ill-posed problem. Conventional approaches primarily concentrate on 2D imagery and fail to fully exploit the capabilities of 3D data representation. 3D point clouds offer a more comprehensive format for representing scenes, as they combine geometric and color information effectively. To this end, in this paper, we introduce Point Intrinsic Net (PoInt-Net), which leverages 3D point cloud data to concurrently estimate albedo and shading maps. The merits of PoInt-Net include the following aspects. First, the model is efficient, achieving consistent performance across point clouds of any size with training only required on small-scale point clouds. Second, it exhibits remarkable robustness; even when trained exclusively on datasets comprising individual objects, PoInt-Net demonstrates strong generalization to unseen objects and scenes. Third, it delivers superior accuracy over conventional 2D approaches, demonstrating enhanced performance across various metrics on different datasets. (Code Released)