Cross-domain generative models based on encoder-decoder AI architectures have attracted much attention in generating realistic images, where domain alignment is crucial for generation accuracy. Domain alignment methods usually deal directly with the initial distribution; however, mismatched or mixed clusters can lead to mode collapse and mixture problems in the decoder, compromising model generalization capabilities. In this work, we innovate a cross-domain alignment and generation model that introduces a canonical latent space representation based on geometric mapping to align the cross-domain latent spaces in a rigorous and precise manner, thus avoiding mode collapse and mixture in the encoder-decoder generation architectures. We name this model GMapLatent. The core of the method is to seamlessly align latent spaces with strict cluster correspondence constraints using the canonical parameterizations of cluster-decorated latent spaces. We first (1) transform the latent space to a canonical parameter domain by composing barycenter translation, optimal transport merging and constrained harmonic mapping, and then (2) compute geometric registration with cluster constraints over the canonical parameter domains. This process realizes a bijective (one-to-one and onto) mapping between newly transformed latent spaces and generates a precise alignment of cluster pairs. Cross-domain generation is then achieved through the aligned latent spaces embedded in the encoder-decoder pipeline. Experiments on gray-scale and color images validate the efficiency, efficacy and applicability of GMapLatent, and demonstrate that the proposed model has superior performance over existing models.
Existing Optimal Transport (OT) methods mainly derive the optimal transport plan/matching under the criterion of transport cost/distance minimization, which may cause incorrect matching in some cases. In many applications, annotating a few matched keypoints across domains is reasonable or even effortless in annotation burden. It is valuable to investigate how to leverage the annotated keypoints to guide the correct matching in OT. In this paper, we propose a novel KeyPoint-Guided model by ReLation preservation (KPG-RL) that searches for the optimal matching (i.e., transport plan) guided by the keypoints in OT. To impose the keypoints in OT, first, we propose a mask-based constraint of the transport plan that preserves the matching of keypoint pairs. Second, we propose to preserve the relation of each data point to the keypoints to guide the matching. The proposed KPG-RL model can be solved by Sinkhorn's algorithm and is applicable even when distributions are supported in different spaces. We further utilize the relation preservation constraint in the Kantorovich Problem and Gromov-Wasserstein model to impose the guidance of keypoints in them. Meanwhile, the proposed KPG-RL model is extended to the partial OT setting. Moreover, we deduce the dual formulation of the KPG-RL model, which is solved using deep learning techniques. Based on the learned transport plan from dual KPG-RL, we propose a novel manifold barycentric projection to transport source data to the target domain. As applications, we apply the proposed KPG-RL model to the heterogeneous domain adaptation and image-to-image translation. Experiments verified the effectiveness of the proposed approach.
1 December 2023 in Journal of Computer-Aided Design & Computer Graphics
Polygon decomposition has been widely employed in the fields of computer graphics, CAD software, and path planning. The presence of self-intersections in polygons introduces errors and inaccuracies in subsequent computations and drawing operations. Decomposing self-intersecting polygons poses a common challenge in CAD applications. Traditional algorithms rely on triangulation techniques. However, this approach yields a substantial number of triangles, significantly elevating both computational and storage complexities. In response to the issue of self-intersecting polygon decomposition, a decomposition algorithm based on region partitioning is proposed. The approach begins by identifying all intersection points within the polygon. Then, a pathfinding technique is employed to traverse the self-intersecting polygon, partitioning it into non-overlapping and self-intersection-free regions. Finally, by discerning the inclusion of each region, internal areas are preserved, while external ones are discarded. This process results in a decomposition of the self-intersecting polygon into non-overlapping simple polygons. Comparative experiments with the GluTess method on large integrated circuit boards reveal that our algorithm improves time efficiency by around 60% and reduces spatial occupancy by about 20%.
18 August 2025 in Journal of Computer-Aided Design & Computer Graphics
N-sided hole filling plays an important role in surface modeling. To address twist compatibility issues, an algorithm based on constrained least square method to fill N-sided holes with NURBS patches is proposed. Firstly, initial cross-boundary derivatives of the surface are generated through linear interpolation of boundary derivatives. Then, a central point is calculated, and this central point along with the midpoints of the boundary lines are used as the first and last control points to construct internal lines, incorporating tangent vectors at the endpoints. Subsequently, normal constraints and twist compatibility conditions are introduced, and the linear interpolation energy of the initial surface boundary derivative are selected as the optimization objective function to generate the cross-boundary derivatives that satisfy boundary compatibility, twist compatibility and ε-G1 continuity, where ε is the angular deviation given by the user. Finally, bicubic blended Coons surfaces are employed to achieve G1 continuous filling for each rectangular region of the N-sided hole. The proposed method solves the numerical instability and surface smoothing problems caused by the small knot vector interval introduced by the Piegl’s algorithm to meet the twist compatibility. The angular deviation between normal vectors of adjacent surfaces is controlled within the user-specified ε angle tolerance. To validate the algorithm’s effectiveness, five representative experimental datasets, each emphasizing distinct aspects of size and curvature, were selected.