A Robust 3D-Based Color Correction Approach for Texture Mapping Applications

Abstract

Texture mapping of 3D models using multiple images often results in textured meshes with unappealing visual artifacts known as texture seams. These artifacts can be more or less visible, depending on the color similarity between the used images. The main goal of this work is to produce textured meshes free of texture seams through a process of color correcting all images of the scene. To accomplish this goal, we propose two contributions to the state-of-the-art of color correction: a pairwise-based methodology, capable of color correcting multiple images from the same scene; the application of 3D information from the scene, namely meshes and point clouds, to build a filtering procedure, in order to produce a more reliable spatial registration between images, thereby increasing the robustness of the color correction procedure. We also present a texture mapping pipeline that receives uncorrected images, an untextured mesh, and point clouds as inputs, producing a final textured mesh and color corrected images as output. Results include a comparison with four other color correction approaches. These show that the proposed approach outperforms all others, both in qualitative and quantitative metrics. The proposed approach enhances the visual quality of textured meshes by eliminating most of the texture seams.

A Sequential Color Correction Approach for Texture Mapping of 3D Meshes

“…Our previous work [ 59 ] has focused on the usage of 3D information from the scene, namely point clouds and 3D mesh, to improve the effectiveness of the color correction procedure, by filtering incorrect correspondences between images from an indoor scenario. These incorrect correspondences are usually due to occlusions and different sizes of the overlapped regions and may lead to poor performance of the color correction algorithm.…”

“…This paper focuses on three components of the pipeline: color correction weighted graph creation ( Section 3.3 ), color correction paths computation ( Section 3.4 ), and sequential pairwise color correction ( Section 3.7 ). Other components, such as pairwise mappings computation and pairwise mappings filtering, have been addressed in detail in our previous work [ 59 ]. In this context, we consider it important for the reader to have an overview of the complete framework, which is why we offer a less detailed explanation of these procedures in Section 3.1 and Section 3.2 .…”

“…As discussed above, our previous work [ 59 ] initially proposed to use 3D meshes to compute correspondences between image pairs. For this reason, the proposed approach starts with the computation of the pairwise mappings that takes as input the RGB images and the 3D mesh.…”

“…However, in situations where the angle between the camera viewpoint (focal axis of the camera) and the normal vector of the face is excessively oblique, or in other words, close to , the impact of the registration error on the accuracy of the pairwise mappings is amplified, increasing the possibility of inaccurate pairwise mappings. In this paper, we leverage previous work [ 59 ], which uses a filtering procedure composed of a combination of 3 filters: z-buffering filtering, depth consistency filtering and camera viewpoint filtering. Each pairwise mapping is only considered valid if it passes the three filtering steps.…”