Place recognition technology endows a SLAM algorithm with the ability to eliminate accumulated errors and to relocalize itself. Existing methods on point cloud-based place recognition often leverage the matching of global descriptors which are lidar-centric. These methods have the following two major defects: place recognition cannot be performed when the distance between the two point clouds is far, and only the rotation angle can be calculated without the offset in the X and Y direction. To solve these two problems, we propose a novel global descriptor, which is built around the Main Object, in this way, descriptors are no longer dependent on the observation position. We analyze the theory that this method can perfectly solve the above two problems, and conduct a lot of experiments in KITTI and some extreme scenarios, which show that our method has obvious advantages over traditional methods.
相關內容
Instance segmentation on point clouds is crucially important for 3D scene understanding. Distance clustering is commonly used in state-of-the-art methods (SOTAs), which is typically effective but does not perform well in segmenting adjacent objects with the same semantic label (especially when they share neighboring points). Due to the uneven distribution of offset points, these existing methods can hardly cluster all instance points. To this end, we design a novel divide and conquer strategy and propose an end-to-end network named PBNet that binarizes each point and clusters them separately to segment instances. PBNet divides offset instance points into two categories: high and low density points (HPs vs.LPs), which are then conquered separately. Adjacent objects can be clearly separated by removing LPs, and then be completed and refined by assigning LPs via a neighbor voting method. To further reduce clustering errors, we develop an iterative merging algorithm based on mean size to aggregate fragment instances. Experiments on ScanNetV2 and S3DIS datasets indicate the superiority of our model. In particular, PBNet achieves so far the best AP50 and AP25 on the ScanNetV2 official benchmark challenge (Validation Set) while demonstrating high efficiency.
Most prior works in perceiving 3D humans from images reason human in isolation without their surroundings. However, humans are constantly interacting with the surrounding objects, thus calling for models that can reason about not only the human but also the object and their interaction. The problem is extremely challenging due to heavy occlusions between humans and objects, diverse interaction types and depth ambiguity. In this paper, we introduce CHORE, a novel method that learns to jointly reconstruct the human and the object from a single RGB image. CHORE takes inspiration from recent advances in implicit surface learning and classical model-based fitting. We compute a neural reconstruction of human and object represented implicitly with two unsigned distance fields, a correspondence field to a parametric body and an object pose field. This allows us to robustly fit a parametric body model and a 3D object template, while reasoning about interactions. Furthermore, prior pixel-aligned implicit learning methods use synthetic data and make assumptions that are not met in the real data. We propose a elegant depth-aware scaling that allows more efficient shape learning on real data. Experiments show that our joint reconstruction learned with the proposed strategy significantly outperforms the SOTA. Our code and models are available at //virtualhumans.mpi-inf.mpg.de/chore
Monocular 3D detection has drawn much attention from the community due to its low cost and setup simplicity. It takes an RGB image as input and predicts 3D boxes in the 3D space. The most challenging sub-task lies in the instance depth estimation. Previous works usually use a direct estimation method. However, in this paper we point out that the instance depth on the RGB image is non-intuitive. It is coupled by visual depth clues and instance attribute clues, making it hard to be directly learned in the network. Therefore, we propose to reformulate the instance depth to the combination of the instance visual surface depth (visual depth) and the instance attribute depth (attribute depth). The visual depth is related to objects' appearances and positions on the image. By contrast, the attribute depth relies on objects' inherent attributes, which are invariant to the object affine transformation on the image. Correspondingly, we decouple the 3D location uncertainty into visual depth uncertainty and attribute depth uncertainty. By combining different types of depths and associated uncertainties, we can obtain the final instance depth. Furthermore, data augmentation in monocular 3D detection is usually limited due to the physical nature, hindering the boost of performance. Based on the proposed instance depth disentanglement strategy, we can alleviate this problem. Evaluated on KITTI, our method achieves new state-of-the-art results, and extensive ablation studies validate the effectiveness of each component in our method. The codes are released at //github.com/SPengLiang/DID-M3D.
Lighting is a determining factor in photography that affects the style, expression of emotion, and even quality of images. Creating or finding satisfying lighting conditions, in reality, is laborious and time-consuming, so it is of great value to develop a technology to manipulate illumination in an image as post-processing. Although previous works have explored techniques based on the physical viewpoint for relighting images, extensive supervisions and prior knowledge are necessary to generate reasonable images, restricting the generalization ability of these works. In contrast, we take the viewpoint of image-to-image translation and implicitly merge ideas of the conventional physical viewpoint. In this paper, we present an Illumination-Aware Network (IAN) which follows the guidance from hierarchical sampling to progressively relight a scene from a single image with high efficiency. In addition, an Illumination-Aware Residual Block (IARB) is designed to approximate the physical rendering process and to extract precise descriptors of light sources for further manipulations. We also introduce a depth-guided geometry encoder for acquiring valuable geometry- and structure-related representations once the depth information is available. Experimental results show that our proposed method produces better quantitative and qualitative relighting results than previous state-of-the-art methods. The code and models are publicly available on //github.com/NK-CS-ZZL/IAN.
Camera-based 3D object detectors are welcome due to their wider deployment and lower price than LiDAR sensors. We first revisit the prior stereo detector DSGN for its stereo volume construction ways for representing both 3D geometry and semantics. We polish the stereo modeling and propose the advanced version, DSGN++, aiming to enhance effective information flow throughout the 2D-to-3D pipeline in three main aspects. First, to effectively lift the 2D information to stereo volume, we propose depth-wise plane sweeping (DPS) that allows denser connections and extracts depth-guided features. Second, for grasping differently spaced features, we present a novel stereo volume -- Dual-view Stereo Volume (DSV) that integrates front-view and top-view features and reconstructs sub-voxel depth in the camera frustum. Third, as the foreground region becomes less dominant in 3D space, we propose a multi-modal data editing strategy -- Stereo-LiDAR Copy-Paste, which ensures cross-modal alignment and improves data efficiency. Without bells and whistles, extensive experiments in various modality setups on the popular KITTI benchmark show that our method consistently outperforms other camera-based 3D detectors for all categories. Code is available at //github.com/chenyilun95/DSGN2.
We present TOCH, a method for refining incorrect 3D hand-object interaction sequences using a data prior. Existing hand trackers, especially those that rely on very few cameras, often produce visually unrealistic results with hand-object intersection or missing contacts. Although correcting such errors requires reasoning about temporal aspects of interaction, most previous works focus on static grasps and contacts. The core of our method are TOCH fields, a novel spatio-temporal representation for modeling correspondences between hands and objects during interaction. TOCH fields are a point-wise, object-centric representation, which encode the hand position relative to the object. Leveraging this novel representation, we learn a latent manifold of plausible TOCH fields with a temporal denoising auto-encoder. Experiments demonstrate that TOCH outperforms state-of-the-art 3D hand-object interaction models, which are limited to static grasps and contacts. More importantly, our method produces smooth interactions even before and after contact. Using a single trained TOCH model, we quantitatively and qualitatively demonstrate its usefulness for correcting erroneous sequences from off-the-shelf RGB/RGB-D hand-object reconstruction methods and transferring grasps across objects.
The objective of this work is to learn an object-centric video representation, with the aim of improving transferability to novel tasks, i.e., tasks different from the pre-training task of action classification. To this end, we introduce a new object-centric video recognition model based on a transformer architecture. The model learns a set of object-centric summary vectors for the video, and uses these vectors to fuse the visual and spatio-temporal trajectory `modalities' of the video clip. We also introduce a novel trajectory contrast loss to further enhance objectness in these summary vectors. With experiments on four datasets -- SomethingSomething-V2, SomethingElse, Action Genome and EpicKitchens -- we show that the object-centric model outperforms prior video representations (both object-agnostic and object-aware), when: (1) classifying actions on unseen objects and unseen environments; (2) low-shot learning to novel classes; (3) linear probe to other downstream tasks; as well as (4) for standard action classification.
Object reconstruction from 3D point clouds has achieved impressive progress in the computer vision and computer graphics research field. However, reconstruction from time-varying point clouds (a.k.a. 4D point clouds) is generally overlooked. In this paper, we propose a new network architecture, namely RFNet-4D, that jointly reconstruct objects and their motion flows from 4D point clouds. The key insight is that simultaneously performing both tasks via learning spatial and temporal features from a sequence of point clouds can leverage individual tasks, leading to improved overall performance. To prove this ability, we design a temporal vector field learning module using unsupervised learning approach for flow estimation, leveraged by supervised learning of spatial structures for object reconstruction. Extensive experiments and analyses on benchmark dataset validated the effectiveness and efficiency of our method. As shown in experimental results, our method achieves state-of-the-art performance on both flow estimation and object reconstruction while performing much faster than existing methods in both training and inference. Our code and data are available at //github.com/hkust-vgd/RFNet-4D
The success of deep learning depends heavily on the availability of large datasets, but in robotic manipulation there are many learning problems for which such datasets do not exist. Collecting these datasets is time-consuming and expensive, and therefore learning from small datasets is an important open problem. Within computer vision, a common approach to a lack of data is data augmentation. Data augmentation is the process of creating additional training examples by modifying existing ones. However, because the types of tasks and data differ, the methods used in computer vision cannot be easily adapted to manipulation. Therefore, we propose a data augmentation method for robotic manipulation. We argue that augmentations should be valid, relevant, and diverse. We use these principles to formalize augmentation as an optimization problem, with the objective function derived from physics and knowledge of the manipulation domain. This method applies rigid body transformations to trajectories of geometric state and action data. We test our method in two scenarios: 1) learning the dynamics of planar pushing of rigid cylinders, and 2) learning a constraint checker for rope manipulation. These two scenarios have different data and label types, yet in both scenarios, training on our augmented data significantly improves performance on downstream tasks. We also show how our augmentation method can be used on real-robot data to enable more data-efficient online learning.
While monocular 3D pose estimation seems to have achieved very accurate results on the public datasets, their generalization ability is largely overlooked. In this work, we perform a systematic evaluation of the existing methods and find that they get notably larger errors when tested on different cameras, human poses and appearance. To address the problem, we introduce VirtualPose, a two-stage learning framework to exploit the hidden "free lunch" specific to this task, i.e. generating infinite number of poses and cameras for training models at no cost. To that end, the first stage transforms images to abstract geometry representations (AGR), and then the second maps them to 3D poses. It addresses the generalization issue from two aspects: (1) the first stage can be trained on diverse 2D datasets to reduce the risk of over-fitting to limited appearance; (2) the second stage can be trained on diverse AGR synthesized from a large number of virtual cameras and poses. It outperforms the SOTA methods without using any paired images and 3D poses from the benchmarks, which paves the way for practical applications. Code is available at //github.com/wkom/VirtualPose.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191