We tackle open-vocabulary 3D scene segmentation tasks by introducing a novel data generation pipeline and training framework. Our work targets three essential aspects required for an effective dataset: precise 3D region segmentation, comprehensive textual descriptions, and sufficient dataset scale. By leveraging state-of-the-art open-vocabulary image segmentation models and region-aware vision-language models (VLM), we develop an automatic pipeline capable of producing high-quality 3D mask-text pairs. Applying this pipeline to multiple 3D scene datasets, we create Mosaic3D-5.6M, a dataset of more than 30K annotated scenes with 5.6M mask-text pairs - significantly larger than existing datasets. Building on these data, we propose Mosaic3D, a 3D visiual foundation model (3D-VFM) combining a 3D encoder trained with contrastive learning and a lightweight mask decoder for open-vocabulary 3D semantic and instance segmentation. Our approach achieves state-of-the-art results on open-vocabulary 3D semantic and instance segmentation benchmarks including ScanNet200, Matterport3D, and ScanNet++, with ablation studies validating the effectiveness of our large-scale training data. https://nvlabs.github.io/Mosaic3D/

Numerical weather prediction (NWP) models are fundamental in meteorology for simulating and forecasting the behavior of various atmospheric variables. The accuracy of precipitation forecasts and the acquisition of sufficient lead time are crucial for preventing hazardous weather events. However, the performance of NWP models is limited by the nonlinear and unpredictable patterns of extreme weather phenomena driven by temporal dynamics. In this regard, we propose a Self-Supervised Learning with Probabilistic Density Labeling (SSLPDL) for estimating rainfall probability by post-processing NWP forecasts. Our post-processing method uses self-supervised learning (SSL) with masked modeling for reconstructing atmospheric physics variables, enabling the model to learn the dependency between variables. The pre-trained encoder is then utilized in transfer learning to a precipitation segmentation task. Furthermore, we introduce a straightforward labeling approach based on probability density to address the class imbalance in extreme weather phenomena like heavy rain events. Experimental results show that SSLPDL surpasses other precipitation forecasting models in regional precipitation post-processing and demonstrates competitive performance in extending forecast lead times. Our code is available at https://github.com/joonha425/SSLPDL.

Learning to assemble geometric shapes into a larger target structure is a pivotal task in various practical applications. In this work, we tackle this problem by establishing local correspondences between point clouds of part shapes in both coarse- and fine-levels. To this end, we introduce Proxy Match Transform (PMT), an approximate high-order feature transform layer that enables reliable matching between mating surfaces of parts while incurring low costs in memory and compute. Building upon PMT, we introduce a new framework, dubbed Proxy Match TransformeR (PMTR), for the geometric assembly task. We evaluate the proposed PMTR on the large-scale 3D geometric shape assembly benchmark dataset of Breaking Bad and demonstrate its superior performance and efficiency compared to state-of-the-art methods. Project page: https://nahyuklee.github.io/pmtr

The recent progress in implicit 3D representation, i.e., Neural Radiance Fields (NeRFs), has made accurate and photorealistic 3D reconstruction possible in a differentiable manner. This new representation can effectively convey the information of hundreds of high-resolution images in one compact format and allows photorealistic synthesis of novel views. In this work, using the variant of NeRF called Plenoxels, we create the first large-scale radiance fields datasets for perception tasks, called the PeRFception, which consists of two parts that incorporate both object-centric and scene-centric scans for classification and segmentation. It shows a significant memory compression rate (96.4\%) from the original dataset, while containing both 2D and 3D information in a unified form. We construct the classification and segmentation models that directly take this radiance fields format as input and also propose a novel augmentation technique to avoid overfitting on backgrounds of images. The code and data are publicly available in "https://postech-cvlab.github.io/PeRFception/".

Point cloud registration is the task of estimating the rigid transformation that aligns a pair of point cloud fragments. We present an efficient and robust framework for pairwise registration of real-world 3D scans, leveraging Hough voting in the 6D transformation parameter space. First, deep geometric features are extracted from a point cloud pair to compute putative correspondences. We then construct a set of triplets of correspondences to cast votes on the 6D Hough space, representing the transformation parameters in sparse tensors. Next, a fully convolutional refinement module is applied to refine the noisy votes. Finally, we identify the consensus among the correspondences from the Hough space, which we use to predict our final transformation parameters. Our method outperforms state-of-the-art methods on 3DMatch and 3DLoMatch benchmarks while achieving comparable performance on KITTI odometry dataset. We further demonstrate the generalizability of our approach by setting a new state-of-the-art on ICL-NUIM dataset, where we integrate our module into a multi-way registration pipeline.

High-dimensional geometric patterns appear in many computer vision problems. In this work, we present high-dimensional convolutional networks for geometric pattern recognition problems that arise in 2D and 3D registration problems. We first propose high-dimensional convolutional networks from 4 to 32 dimensions and analyze the geometric pattern recognition capacity in high-dimensional linear regression problems. Next, we show that the 3D correspondences form hyper-surface in a 6-dimensional space and validate our network on 3D registration problems. Finally, we use image correspondences, which form a 4-dimensional hyper-conic section, and show that the high-dimensional convolutional networks are on par with many state-of-the-art multi-layered perceptrons.

Class cohesion is a design quality that has a great impact on posterior development and maintenance of object-oriented systems. There have been much works on measuring class cohesion based on internal class relationships such as method-attribute referencing and internal method invocations. However, object-oriented systems involve many external class relationships carrying a functional context which should be also taken into account in measuring class cohesion. However, the existing works using internal relationships cannot measure cohesion of classes involving many external relationships. In this paper, we present a novel metric for measuring class cohesion based on both internal and external class relationships through an iterative process. We evaluate the metric by applying it to five applications — ArgoUML, jEdit, JHotDraw, JMeter, and Refactor, and validate the metric theoretically and statistically. The theoretical validation shows that the metric exhibits the required properties for cohesion metrics and the statistical validation shows that the metric captures a new aspect of class cohesion that has not been addressed by the existing works.

Cohesion significantly impacts on posterior development and maintenance. A well-designed module should be coherent with a single responsibility and less propagative for changes to other modules. In object-oriented development, classes are often modularized based on their relationships to localize change impact. There has been much work on grouping classes using direct class interactions and similarity of lexical information of classes. However, there has been little attention to indirect class relationships which are significant in defining functional semantics in object-oriented systems. In this paper, we present an approach for grouping classes using both direct and indirect class relationships. In the approach, we define a set of metrics for measuring couplings based on indirect class relationships and use them to group classes with the aim of localizing the impact of changes. We evaluate the approach using three applications - JHotDraw, JMeter, and ArgoUML and compare the results with existing work. The evaluation shows that the proposed approach reduces 17.2% and 22.2% of the number of packages that are affected by changes over the existing works.

In this paper, we propose a method to estimate the read failure rate of a static random access memory (SRAM) cell. Conventional read stability metrics cannot accurately estimate the read failure probability as technology scales down, because some metrics cannot characterize read stability and others can no longer be approximated to a known distribution. We first introduce a one-sided static noise margin (OSNM), whose lower tail region follows a Gaussian distribution, and also propose a Gaussian-tail-fitting method that properly models the distribution of the OSNM at the tail region. It is demonstrated that the OSNM can accurately estimate the SRAM cell yield using the proposed Gaussian-tail-fitting method.