Monocular 3D object detection aims to localize 3D bounding boxes in an input single 2D image. It is a highly challenging problem and remains open, especially when no extra information (e.g., depth, lidar and/or multi-frames) can be leveraged in training and/or inference. This paper proposes a simple yet effective formulation for monocular 3D object detection without exploiting any extra information. It presents the MonoCon method which learns Monocular Contexts, as auxiliary tasks in training, to help monocular 3D object detection. The key idea is that with the annotated 3D bounding boxes of objects in an image, there is a rich set of well-posed projected 2D supervision signals available in training, such as the projected corner keypoints and their associated offset vectors with respect to the center of 2D bounding box, which should be exploited as auxiliary tasks in training. The proposed MonoCon is motivated by the Cram\`er–Wold theorem in measure theory at a high level. In experiments, the proposed MonoCon is tested in the KITTI benchmark (car, pedestrian and cyclist). It outperforms all prior arts in the leaderboard on the car category and obtains comparable performance on pedestrian and cyclist in terms of accuracy. Thanks to the simple design, the proposed MonoCon method obtains the fastest inference speed with 38.7 fps in comparisons.
more »
« less
This content will become publicly available on October 1, 2024
Monocular 3D Object Detection with Bounding Box Denoising in 3D by Perceiver
The main challenge of monocular 3D object detection is the accurate localization of 3D center. Motivated by a new and strong observation that this challenge can be remedied by a 3D-space local-grid search scheme in an ideal case, we propose a stage-wise approach, which combines the information flow from 2D-to-3D (3D bounding box proposal generation with a single 2D image) and 3D-to-2D (proposal verification by denoising with 3D-to-2D contexts) in a topdown manner. Specifically, we first obtain initial proposals from off-the-shelf backbone monocular 3D detectors. Then, we generate a 3D anchor space by local-grid sampling from the initial proposals. Finally, we perform 3D bounding box denoising at the 3D-to-2D proposal verification stage. To effectively learn discriminative features for denoising highly overlapped proposals, this paper presents a method of using the Perceiver I/O model [20] to fuse the 3D-to-2D
geometric information and the 2D appearance information. With the encoded latent representation of a proposal, the verification head is implemented with a self-attention module. Our method, named as MonoXiver, is generic and can be easily adapted to any backbone monocular 3D detectors. Experimental results on the well-established KITTI dataset and the challenging large-scale Waymo dataset show that MonoXiver consistently achieves improvement with limited computation overhead.
more »
« less
- NSF-PAR ID:
- 10468300
- Publisher / Repository:
- The Computer Vision Foundation (CVF)
- Date Published:
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
null (Ed.)Three-dimensional objects are commonly represented as 3D boxes in a point-cloud. This representation mimics the well-studied image-based 2D bounding-box detection but comes with additional challenges. Objects in a 3D world do not follow any particular orientation, and box-based detectors have difficulties enumerating all orientations or fitting an axis-aligned bounding box to rotated objects. In this paper, we instead propose to represent, detect, and track 3D objects as points. Our framework, CenterPoint, first detects centers of objects using a keypoint detector and regresses to other attributes, including 3D size, 3D orientation, and velocity. In a second stage, it refines these estimates using additional point features on the object. In CenterPoint, 3D object tracking simplifies to greedy closest-point matching. The resulting detection and tracking algorithm is simple, efficient, and effective. CenterPoint achieved state-of-the-art performance on the nuScenes benchmark for both 3D detection and tracking, with 65.5 NDS and 63.8 AMOTA for a single model. On the Waymo Open Dataset, CenterPoint outperforms all previous single model methods by a large margin and ranks first among all Lidar-only submissions.more » « less
-
more » « less
In this paper we present Sniffer Faster R-CNN++, an efficient Camera-LiDAR late fusion network for low complexity and accurate object detection in autonomous driving scenarios. The proposed detection network architecture operates on output candidates of any 3D detector and proposals from regional proposal network of any 2D detector to generate final prediction results. In comparison to the single modality object detection approaches, fusion based methods in many instances suffer from dissimilar data integration difficulties. On one hand, fusion based network models are complicated in nature and on the other hand they require large computational overhead and resources, processing pipelines for training and inference specially, the early fusion and deep fusion approaches. As such, we devise a late fusion network that in-cooperates pre-trained, single-modality detectors without change, performing association only at the detection level. In addition to this, lidar based method fail to detect distant object due to its sparse nature so we devise proposal refinement algorithm to jointly optimize detection candidates and assist detection for distant objects. Extensive experiments on both the 3D and 2D detection benchmark of challenging KITTI dataset illustrate that our proposed network architecture significantly improves the detection accuracy, accelerating the detection speed.
-
more » « less
Integral imaging (InIm) is useful for passive ranging and 3D visualization of partially-occluded objects. We consider 3D object localization within a scene and in occlusions. 2D localization can be achieved using machine learning and non-machine learning-based techniques. These techniques aim to provide a 2D bounding box around each one of the objects of interest. A recent study uses InIm for the 3D reconstruction of the scene with occlusions and utilizes mutual information (MI) between the bounding box in this 3D reconstructed scene and the corresponding bounding box in the central elemental image to achieve passive depth estimation of partially occluded objects. Here, we improve upon this InIm method by using Bayesian optimization to minimize the number of required 3D scene reconstructions. We evaluate the performance of the proposed approach by analyzing different kernel functions, acquisition functions, and parameter estimation algorithms for Bayesian optimization-based inference for simultaneous depth estimation of objects and occlusion. In our optical experiments, mutual-information-based depth estimation with Bayesian optimization achieves depth estimation with a handful of 3D reconstructions. To the best of our knowledge, this is the first report to use Bayesian optimization for mutual information-based InIm depth estimation.
-
Detecting small objects (e.g., manhole covers, license plates, and roadside milestones) in urban images is a long-standing challenge mainly due to the scale of small object and background clutter. Although convolution neural network (CNN)-based methods have made significant progress and achieved impressive results in generic object detection, the problem of small object detection remains unsolved. To address this challenge, in this study we developed an end-to-end network architecture that has three significant characteristics compared to previous works. First, we designed a backbone network module, namely Reduced Downsampling Network (RD-Net), to extract informative feature representations with high spatial resolutions and preserve local information for small objects. Second, we introduced an Adjustable Sample Selection (ADSS) module which frees the Intersection-over-Union (IoU) threshold hyperparameters and defines positive and negative training samples based on statistical characteristics between generated anchors and ground reference bounding boxes. Third, we incorporated the generalized Intersection-over-Union (GIoU) loss for bounding box regression, which efficiently bridges the gap between distance-based optimization loss and area-based evaluation metrics. We demonstrated the effectiveness of our method by performing extensive experiments on the public Urban Element Detection (UED) dataset acquired by Mobile Mapping Systems (MMS). The Average Precision (AP) of the proposed method was 81.71%, representing an improvement of 1.2% compared with the popular detection framework Faster R-CNN.more » « less
