Outlier removal is a critical part of feature-based point cloud registration. In this paper, we revisit the model generation and selection of the classic RANSAC approach for fast and robust point cloud registration. For the model generation, we propose a second-order spatial compatibility (SC $^{2}$ ) measure to compute the similarity between correspondences. It takes into account global compatibility instead of local consistency, allowing for more distinctive clustering between inliers and outliers at an early stage. The proposed measure promises to find a certain number of outlier-free consensus sets using fewer samplings, making the model generation more efficient. For the model selection, we propose a new Feature and Spatial consistency constrained Truncated Chamfer Distance (FS-TCD) metric for evaluating the generated models. It considers the alignment quality, the feature matching properness, and the spatial consistency constraint simultaneously, enabling the correct model to be selected even when the inlier rate of the putative correspondence set is extremely low. Extensive experiments are carried out to investigate the performance of our method. In addition, we also experimentally prove that the proposed SC $^{2}$ measure and the FS-TCD metric are general and can be easily plugged into deep learning based frameworks. The code will be available at https://github.com/ZhiChen902/SC2-PCR-plusplus .

Point-cloud registration is a fundamental task in computer vision. However, most point clouds are partially overlapping, corrupted by noise and comprised of indistinguishable surfaces, especially for complexly distributed outdoor LiDAR point clouds, which makes registration challenging. In this paper, we propose a multi-scale features-based point cloud registration network named MSPR-Net for large-scale outdoor LiDAR point cloud registration. The main motivation of the proposed MSPR-Net is that the features of two keypoints from a true correspondence must match in different scales. From this point of view, we first utilize a multi-scale backbone to extract the multi-scale features of the keypoints. Next, we propose a bilateral outlier removal strategy to remove the potential outliers in the keypoints based on the multi-scale features. Finally, a coarse-to-fine registration way is applied to exploit the information both in feature and spatial space. Extensive experiments conducted on two large-scale outdoor LiDAR point cloud datasets demonstrate that MSPR-Net achieves state-of-the-art performance.

Human mesh recovery aims to estimate all human meshes within a given image. In this article, we propose an Instance-aware Multi-person 3D Human Mesh Recovery (InstaHMR) network based on the one-stage framework. Compared to former one-stage methods, instance-aware single person feature is exploited to represent more accurate human mesh. Specifically, we propose the Contextual Instance Guidance (CIG) module which generates instance-aware single person feature by leveraging spatial and channel attention operations. In this way, it preserves more instance-specific information compared to the pixel-level feature used in some existing one-stage methods. Besides, we further introduce two auxiliary losses for better mesh recovery, namely the Human Triplet Planes (HTP) loss and the T-pose Shape (TS) loss. The HTP loss encourages the model to capture subtle differences in human joint positions, while the TS loss facilitates the learning of abstract shape parameters. By incorporating these advancements, our model achieves state-of-the-art results on four multi-person datasets.

Homography estimation is a fundamental task in many computer vision applications. Recently, deep learning-based methods have shown to be favorable for this problem when compared to traditional methods. They typically transform the homography estimation into regressing 4 point offsets which can be used to calculate a homography matrix. However, these methods usually use an implicit network to directly regress the offsets and completely ignores the matching information. The implicit regression network could be easy to cause overfitting and has limited representation ability to deal with image details. To address these problems, we propose an explicit dense scene matching strategy which introduces the matching information to establish a more rational connection between images and offsets. In this design, the network is only required to learn for simple feature encoding and information interaction tasks, which significantly reduces the dependence on representation ability of the network. As a result, the proposed explicit strategy can lead to better performance in aligning image details. Based on the dense scene matching, we propose a coarse-to-fine network consisting of three sub-models to progressively align images. Experimental results reveal that our method outperforms the state-of-the-art single-homography methods and multi-homography methods.

Transformer architecture has shown significant potential in various visual tasks, including point cloud registration. Positional encoding, as an order-aware module, plays a crucial role in Transformer framework. In this paper, we propose OIF-PCR++, a conditional positional encoding (CPE) method for point cloud registration. The core CPE module utilizes length and vector encoding at different stages, conditioned on the relative pose states between the point clouds to be registered. As a result, it progressively alleviates feature ambiguity through the incorporation of geometric cues. Building upon CPE, we introduce an iterative positional encoding optimization pipeline comprising two stages: 1) We find one correspondence via a differentiable optimal transport layer, and use it to encode length information into point cloud features, enhancing spatial consistency across different reference frames. 2) We apply a progressive direction alignment strategy to achieve rough alignment between paired point clouds, and then gradually incorporate direction information with the aid of this alignment, further enhancing feature distinctiveness and reducing feature ambiguity. Through this iterative optimization process, length and direction information are effectively integrated to achieve consistent and distinctive positional encoding, enabling the learning of discriminative point cloud features. Additionally, we present an inlier propagation mechanism that harmoniously integrates consistent geometric information for positional encoding. The proposed method is highly efficient, introducing marginal computational overhead while significantly improving feature distinguishability. Extensive experiments demonstrate superior performance over state-of-the-art methods on indoor, outdoor, object-level, and multi-way benchmarks, as well as strong generalization to complex real-world scenarios.