Feature Preserving Mesh Denoising Based on Graph Spectral Processing

Human-airway surface mesh smoothing based on graph convolutional neural networks

BACKGROUND AND OBJECTIVE

A detailed representation of the airway geometry in the respiratory system is critical for predicting precise airflow and pressure behaviors in computed tomography (CT)-image-based computational fluid dynamics (CFD). The CT-image-based geometry often contains artifacts, noise, and discontinuities due to the so-called stair step effect. Hence, an advanced surface smoothing is necessary. The existing smoothing methods based on the Laplacian operator drastically shrink airway geometries, resulting in the loss of information related to smaller branches. This study aims to introduce an unsupervised airway-mesh-smoothing learning (AMSL) method that preserves the original geometry of the three-dimensional (3D) airway for accurate CT-image-based CFD simulations.

METHOD

The AMSL method jointly trains two graph convolutional neural networks (GCNNs) defined on airway meshes to filter vertex positions and face normal vectors. In addition, it regularizes a combination of loss functions such as reproducibility, smoothness and consistency of vertex positions, and normal vectors. The AMSL adopts the concept of a deep mesh prior model, and it determines the self-similarity for mesh restoration without using a large dataset for training. Images of the airways of 20 subjects were smoothed by the AMSL method, and among them, the data of two subjects were used for the CFD simulations to assess the effect of airway smoothing on flow properties.

RESULTS

In 18 of 20 benchmark problems, the proposed smoothing method delivered better results compared with the conventional or state-of-the-art deep learning methods. Unlike the traditional smoothing, the AMSL successfully constructed 20 smoothed airways with airway diameters that were consistent with the original CT images. Besides, CFD simulations with the airways obtained by the AMSL method showed much smaller pressure drop and wall shear stress than the results obtained by the traditional method.

CONCLUSIONS

The airway model constructed by the AMSL method reproduces branch diameters accurately without any shrinkage, especially in the case of smaller airways. The accurate estimation of airway geometry using a smoothing method is critical for estimating flow properties in CFD simulations.

Mesh Total Generalized Variation for Denoising

Recent studies have shown that the Total Generalized Variation (TGV) is highly effective in preserving sharp features as well as smooth transition variations for image processing tasks. However, currently there is no existing work that is suitable for applying TGV to 3D data, in particular, triangular meshes. In this article, we develop a novel framework for discretizing second-order TGV on triangular meshes. Further, we propose a TGV-based variational method for the denoising of face normal fields on triangular meshes. The TGV regularizer in our method is composed of a first-order term and a second-order term, which are automatically balanced. The first-order term allows our TGV regularizer to locate and preserve sharp features, while the second-order term allows our regularizer to recognize and recover smoothly curved regions. To solve the optimization problem, we introduce an efficient iterative algorithm based on variable-splitting and augmented Lagrangian method. Extensive results and comparisons on synthetic and real scanning data validate that the proposed method outperforms the state-of-the-art visually and numerically.

GDR-Net: A Geometric Detail Recovering Network for 3D Scanned Objects

This article addresses the problem of mesh super-resolution such that the geometry details which are not well represented in the low-resolution models can be recovered and well represented in the generated high-quality models. The main challenges of this problem are the nonregularity of 3D mesh representation and the high complexity of 3D shapes. We propose a deep neural network called GDR-Net to solve this ill-posed problem, which resolves the two challenges simultaneously. First, to overcome the nonregularity, we regress a displacement in radial basis function parameter space instead of the vertex-wise coordinates in the euclidean space. Second, to overcome the high complexity, we apply the detail recovery process to small surface patches extracted from the input surface and obtain the overall high-quality mesh by fusing the refined surface patches. To train the network, we constructed a dataset composed of both real-world and synthetic scanned models, including high/low-quality pairs. Our experimental results demonstrate that GDR-Net works well for general models and outperforms previous methods for recovering geometric details.

Mesh Denoising With Facet Graph Convolutions

We examine the problem of mesh denoising, which consists of removing noise from corrupted 3D meshes while preserving existing geometric features. Most mesh denoising methods require a lot of mesh-specific parameter fine-tuning, to account for specific features and noise types. In recent years, data-driven methods have demonstrated their robustness and effectiveness with respect to noise and feature properties on a wide variety of geometry and image problems. Most existing mesh denoising methods still use hand-crafted features, and locally denoise facets rather than examine the mesh globally. In this work, we propose the use of a fully end-to-end learning strategy based on graph convolutions, where meaningful features are learned directly by our network. It operates on a graph of facets, directly on the existing topology of the mesh, without resampling, and follows a multi-scale design to extract geometric features at different resolution levels. Similar to most recent pipelines, given a noisy mesh, we first denoise face normals with our novel approach, then update vertex positions accordingly. Our method performs significantly better than the current state-of-the-art learning-based methods. Additionally, we show that it can be trained on noisy data, without explicit correspondence between noisy and ground-truth facets. We also propose a multi-scale denoising strategy, better suited to correct noise with a low spatial frequency.

DNF-Net: A Deep Normal Filtering Network for Mesh Denoising

This article presents a deep normal filtering network, called DNF-Net, for mesh denoising. To better capture local geometry, our network processes the mesh in terms of local patches extracted from the mesh. Overall, DNF-Net is an end-to-end network that takes patches of facet normals as inputs and directly outputs the corresponding denoised facet normals of the patches. In this way, we can reconstruct the geometry from the denoised normals with feature preservation. Besides the overall network architecture, our contributions include a novel multi-scale feature embedding unit, a residual learning strategy to remove noise, and a deeply-supervised joint loss function. Compared with the recent data-driven works on mesh denoising, DNF-Net does not require manual input to extract features and better utilizes the training data to enhance its denoising performance. Finally, we present comprehensive experiments to evaluate our method and demonstrate its superiority over the state of the art on both synthetic and real-scanned meshes.

Graph-Based Feature-Preserving Mesh Normal Filtering

Distinguishing between geometric features and noise is of paramount importance for mesh denoising. In this paper, a graph-based feature-preserving mesh normal filtering scheme is proposed, which includes two stages: graph-based feature detection and feature-aware guided normal filtering. In the first stage, faces in the input noisy mesh are represented by patches, which are then modelled as weighted graphs. In this way, feature detection can be cast as a graph-cut problem. Subsequently, an iterative normalized cut algorithm is applied on each patch to separate the patch into smooth regions according to the detected features. In the second stage, a feature-aware guidance normal is constructed for each face, and guided normal filtering is applied to achieve robust feature-preserving mesh denoising. The results of experiments on synthetic and real scanned models indicate that the proposed scheme outperforms state-of-the-art mesh denoising works in terms of both objective and subjective evaluations.

Spectral Processing for Denoising and Compression of 3D Meshes Using Dynamic Orthogonal Iterations

Recently, spectral methods have been extensively used in the processing of 3D meshes. They usually take advantage of some unique properties that the eigenvalues and the eigenvectors of the decomposed Laplacian matrix have. However, despite their superior behavior and performance, they suffer from computational complexity, especially while the number of vertices of the model increases. In this work, we suggest the use of a fast and efficient spectral processing approach applied to dense static and dynamic 3D meshes, which can be ideally suited for real-time denoising and compression applications. To increase the computational efficiency of the method, we exploit potential spectral coherence between adjacent parts of a mesh and then we apply an orthogonal iteration approach for the tracking of the graph Laplacian eigenspaces. Additionally, we present a dynamic version that automatically identifies the optimal subspace size that satisfies a given reconstruction quality threshold. In this way, we overcome the problem of the perceptual distortions, due to the fixed number of subspace sizes that is used for all the separated parts individually. Extensive simulations carried out using different 3D models in different use cases (i.e., compression and denoising), showed that the proposed approach is very fast, especially in comparison with the SVD based spectral processing approaches, while at the same time the quality of the reconstructed models is of similar or even better reconstruction quality. The experimental analysis also showed that the proposed approach could also be used by other denoising methods as a preprocessing step, in order to optimize the reconstruction quality of their results and decrease their computational complexity since they need fewer iterations to converge.