Object detection with multimodal inputs can improve many safety-critical systems such as autonomous vehicles (AVs). Motivated by AVs that operate in both day and night, we study multimodal object detection with RGB and thermal cameras, since the latter provides much stronger object signatures under poor illumination. We explore strategies for fusing information from different modalities. Our key contribution is a probabilistic ensembling technique, ProbEn, a simple non-learned method that fuses together detections from multi-modalities. We derive ProbEn from Bayes’ rule and first principles that assume conditional independence across modalities. Through probabilistic marginalization, ProbEn elegantly handles missing modalities when detectors do not fire on the same object. Importantly, ProbEn also notably improves multimodal detection even when the conditional independence assumption does not hold, e.g., fusing outputs from other fusion methods (both off-the-shelf and trained in-house). We validate ProbEn on two benchmarks containing both aligned (KAIST) and unaligned (FLIR) multimodal images, showing that ProbEn outperforms prior work by more than {\bf 13\%} in relative performance!
论文检索
输入标题、作者或关键词，从 1,645 篇学术成果中精准定位
Real-world object detection models should be cheap and accurate. Knowledge distillation (KD) can boost the accuracy of a small, light detection model by leveraging useful information from a larger teacher model. However, a key challenge is identifying the most informative features produced by the teacher for distillation. In this work, we show that only a very small fraction of features within a ground-truth bounding box are responsible for a teacher’s high detection performance. Based on this, we propose Prediction-Guided Distillation (PGD), which focuses distillation on these key predictive regions of the teacher and yields considerable gains in performance over many existing KD baselines. In addition, we propose an adaptive weighting scheme over the key regions to smooth out their influence and achieve even better performance. Our proposed approach outperforms current state-of-the-art KD baselines on a variety of advanced one-stage detection architectures. Specifically, on the COCO dataset, our method achieves between +3.1% and +4.6% AP improvement using ResNet-101 and ResNet-50 as the teacher and student backbones, respectively. On the CrowdHuman dataset, we achieve +3.2% and +2.0% improvements in MR and AP, also using these backbones. Our code is available at https://github.com/ChenhongyiYang/PGD.
Open-vocabulary object detection, which is concerned with the problem of detecting novel objects guided by natural language, has gained increasing attention from the community. Ideally, we would like to extend an open-vocabulary detector such that it can produce bounding box predictions based on user inputs in form of either natural language or exemplar image. This offers great flexibility and user experience for human-computer interaction. To this end, we propose a novel open-vocabulary detector based on DETR---hence the name OV-DETR---which, once trained, can detect any object given its class name or an exemplar image. The biggest challenge of turning DETR into an open-vocabulary detector is that it is impossible to calculate the classification cost matrix of novel classes without access to their labeled images. To overcome this challenge, we formulate the learning objective as a binary matching one between input queries (class name or exemplar image) and the corresponding objects, which learns useful correspondence to generalize to unseen queries during testing. For training, we choose to condition the Transformer decoder on the input embeddings obtained from a pre-trained vision-language model like CLIP, in order to enable matching for both text and image queries. With extensive experiments on LVIS and COCO datasets, we demonstrate that our OV-DETR---the first end-to-end Transformer-based open-vocabulary detector---achieves non-trivial improvements over current state of the arts.
Detection transformers have achieved competitive performance on the sample-rich COCO dataset. However, we show most of them suffer from significant performance drops on small-size datasets, like Cityscapes. In other words, the detection transformers are generally data-hungry. To tackle this problem, we empirically analyze the factors that affect data efficiency, through a step-by-step transition from a data-efficient RCNN variant to the representative DETR. The empirical results suggest that sparse feature sampling from local image areas holds the key. Based on this observation, we alleviate the data-hungry issue of existing detection transformers by simply alternating how key and value sequences are constructed in the cross-attention layer, with minimum modifications to the original models. Besides, we introduce a simple yet effective label augmentation method to provide richer supervision and improve data efficiency. Experiments show that our method can be readily applied to different detection transformers and improve their performance on both small-size and sample-rich datasets.
We aim to improve the performance of regressing hand keypoints and segmenting pixel-level hand masks under new imaging conditions (e.g., outdoors) when we only have labeled images taken under very different conditions (e.g., indoors). In the real world, it is important that the model trained for both tasks works under various imaging conditions. However, their variation covered by existing labeled hand datasets is limited. Thus, it is necessary to adapt the model trained on the labeled images (source) to unlabeled images (target) with unseen imaging conditions. While self-training domain adaptation methods (i.e., learning from the unlabeled target images in a self-supervised manner) have been developed for both tasks, their training may degrade performance when the predictions on the target images are noisy. To avoid this, it is crucial to assign a low importance (confidence) weight to the noisy predictions during self-training. In this paper, we propose to utilize the divergence of two predictions to estimate the confidence of the target image for both tasks. These predictions are given from two separate networks, and their divergence helps identify the noisy predictions. To integrate our proposed confidence estimation into self-training, we propose a teacher-student framework where the two networks (teachers) provide supervision to a network (student) for self-training, and the teachers are learned from the student by knowledge distillation. Our experiments show its superiority over state-of-the-art methods in adaptation settings with different lighting, grasping objects, backgrounds, and camera viewpoints. Our method improves by 4% the multi-task score on HO3D compared to the latest adversarial adaptation method. We also validate our method on Ego4D, egocentric videos with rapid changes in imaging conditions outdoors.
Object detection using single point supervision has received increasing attention over the years. However, the performance gap between point supervised object detection (PSOD) and bounding box supervised detection remains large. In this paper, we attribute such a large performance gap to the failure of generating high-quality proposal bags which are crucial for multiple instance learning (MIL). To address this problem, we introduce a lightweight alternative to the off-the-shelf proposal (OTSP) method and thereby create the Point-to-Box Network (P2BNet), which can construct a inter-objects balanced proposal bag by generating proposals in an anchor-like way. By fully investigating the accurate position information, P2BNet further constructs an instance-level bag, avoiding the mixture of multiple objects. Finally, a coarse-to-fine policy in a cascade fashion is utilized to improve the IoU between proposals and ground-truth (GT). Benefiting from these strategies, P2BNet is able to produce high-quality instance-level bags for object detection. P2BNet improves the mean average precision (AP) by more than 50% relative to the previous best PSOD method on the MS COCO dataset. It also demonstrates the great potential to bridge the performance gap between point supervised and bounding-box supervised detectors. The code will be released at github.com/ucas-vg/P2BNet.
To date, the most powerful semi-supervised object detectors (SS-OD) are based on pseudo-boxes, which need a sequence of post-processing with fine-tuned hyper-parameters. In this work, we propose replacing the sparse pseudo-boxes with the dense prediction as a united and straightforward form of pseudo-label. Compared to the pseudo-boxes, our Dense Pseudo-Label (DPL) does not involve any post-processing method, thus retaining richer information. We also introduce a region selection technique to highlight the key information while suppressing the noise carried by dense labels. We name our proposed SS-OD algorithm that leverages the DPL as Dense Teacher. On COCO and VOC, Dense Teacher shows superior performance under various settings compared with the pseudo-box-based methods. Code will be available.
Category-Level 6D Object Pose and Size Estimation Using Self-Supervised Deep Prior Deformation Networks
PDF ↗It is difficult to precisely annotate object instances and their semantics in 3D space, and as such, synthetic data are extensively used for these tasks, e.g., category-level 6D object pose and size estimation. However, the easy annotations in synthetic domains bring the downside effect of synthetic-to-real (Sim2Real) domain gap. In this work, we aim to address this issue in the task setting of Sim2Real, unsupervised domain adaptation for category-level 6D object pose and size estimation. We propose a method that is built upon a novel Deep Prior Deformation Network, shortened as DPDN. DPDN learns to deform features of categorical shape priors to match those of object observations, and is thus able to establish deep correspondence in the feature space for direct regression of object poses and sizes. To reduce the Sim2Real domain gap, we formulate a novel self-supervised objective upon DPDN via consistency learning; more specifically, we apply two rigid transformations to each object observation in parallel, and feed them into DPDN respectively to yield dual sets of predictions; on top of the parallel learning, an inter-consistency term is employed to keep cross consistency between dual predictions for improving the sensitivity of DPDN to pose changes, while individual intra-consistency ones are used to enforce self-adaptation within each learning itself. We train DPDN on both training sets of the synthetic CAMERA25 and real-world REAL275 datasets; our results outperform the existing methods on REAL275 test set under both the unsupervised and supervised settings. Ablation studies also verify the efficacy of our designs. Our code is released publicly at https://github.com/JiehongLin/Self-DPDN.
BEVFormer: Learning Bird’s-Eye-View Representation from Multi-Camera Images via Spatiotemporal Transformers
PDF ↗3D visual perception tasks, including 3D detection and map segmentation based on multi-camera images, are essential for autonomous driving systems. In this work, we present a new framework termed BEVFormer, which learns unified BEV representations with spatiotemporal transformers to support multiple autonomous driving perception tasks. In a nutshell, BEVFormer exploits both spatial and temporal information by interacting with spatial and temporal space through predefined grid-shaped BEV queries. To aggregate spatial information, we design spatial cross-attention that each BEV query extracts the spatial features from the regions of interest across camera views. For temporal information, we propose temporal self-attention to recurrently fuse the history BEV information. Our approach achieves the new state-of-the-art 56.9% in terms of NDS metric on the nuScenes test set, which is 9.0 points higher than previous best arts and on par with the performance of LiDAR-based baselines. We further show that BEVFormer remarkably improves the accuracy of velocity estimation and recall of objects under low visibility conditions. The code will be released at https://github.com/zhiqi-li/BEVFormer.
Recent years have witnessed a trend of applying context frames to boost the performance of object detection as video object detection. Existing methods usually aggregate features at one stroke to enhance the feature. These methods, however, usually lack spatial information from neighboring frames and suffer from insufficient feature aggregation. To address the issues, we perform a progressive way to introduce both temporal information and spatial information for an integrated enhancement. The temporal information is introduced by the temporal feature aggregation model (TFAM), by conducting an attention mechanism between the context frames and the target frame (i.e., the frame to be detected). Meanwhile, we employ a Spatial Transition Awareness Model (STAM) to convey the location transition information between each context frame and target frame. Built upon a transformer-based detector DETR, our PTSEFormer also follows an end-to-end fashion to avoid heavy post-processing procedures while achieving 88.1% mAP on the ImageNet VID dataset. Codes are available at https://github.com/Hon-Wong/PTSEFormer.
The success of monocular 3D object detection highly relies on considerable labeled data, which is costly to obtain. To alleviate the annotation effort, we propose MVC-MonoDet, the first semi-supervised training framework that improves Monocular 3D object detection by enforcing multi-view consistency. In particular, a box-level regularization and an object-level regularization are designed to enforce the consistency of 3D bounding box predictions of the detection model across unlabeled multi-view data (stereo or video). The box-level regularizer requires the model to consistently estimate 3D boxes in different views so that the model can learn cross-view invariant features for 3D detection. The object-level regularizer employs an object-wise photometric consistency loss that mitigates 3D box estimation error through structure from motion (SFM). A key innovation in our approach to effectively utilize these consistency losses from multi-view data is a novel relative depth module that replaces the standard depth module in vanilla SFM. This technique allows the depth estimation to be coupled with the estimated 3D bounding boxes, so that the derivative of consistency regularization can be used to directly optimize the estimated 3D bounding boxes using unlabeled data. We show that the proposed semi-supervised learning techniques effectively improve the performance of 3D detection on the KITTI and nuScenes datasets. We also demonstrate that the framework is flexible and can be adapted to both stereo and video data.
Large-scale object detection and instance segmentation faces a severe data imbalance. The finer-grained object classes become, the less frequent they appear in our datasets. However at test-time, we expect a detector that performs well for all classes and not just the most frequent ones. In this paper, we provide a theoretical understanding of the long-trail detection problem. We show how the commonly used mean average precision evaluation metric on an unknown test-set is bound by a margin-based binary classification error on a long-tailed object-detection training set. We optimize margin-based binary classification error with a novel surrogate objective called Effective Class-Margin Loss (ECM). The ECM loss is simple, theoretically well-motivated, and outperforms other heuristic counterparts on LVIS v1 benchmark over a wide range of architecture and detectors. Code is available at https://github.com/janghyuncho/ECM-Loss.
Accurate and reliable 3D detection is vital for many applications including autonomous driving vehicles and service robots. In this paper, we present a flexible and high-performance 3D detection frame-work, named MPPNet, for 3D temporal object detection with point cloud sequences. We propose a novel three-hierarchy framework with proxy points for multi-frame feature encoding and interactions to achieve better detection. The three hierarchies conduct per-frame feature encoding, short-clip feature fusion, and whole-sequence feature aggregation, respectively. To enable processing long-sequence point clouds with reasonable computational resources, intra-group feature mixing and inter-group feature attention are proposed to form the second and third feature encoding hierarchies, which are recurrently applied for aggregating multi-frame trajectory features. The proxy points not only act as consistent object representations for each frame, but also serves as the courier to facilitate feature interaction between frames. The experiments on large Waymo Open dataset show that our approach outperforms state-of-the-art methods with large margins when applied to both short (e.g., 4-frame) and long (e.g., 16-frame) point cloud sequences. Code will be publicly available at https://github.com/open-mmlab/OpenPCDet.
Two-stage detectors have gained much popularity in 3D object detection. Most two-stage 3D detectors utilize grid points, voxel grids, or sampled keypoints for RoI feature extraction in the second stage. Such methods, however, are inefficient in handling unevenly distributed and sparse outdoor points. This paper solves this problem in three aspects. 1) Dynamic Point Aggregation. We propose the patch search to quickly search points in a local region for each 3D proposal. The dynamic farthest voxel sampling is then applied to evenly sample the points. Especially, the voxel size varies along the distance to accommodate the uneven distribution of points. 2) RoI-graph Pooling. We build local graphs on the sampled points to better model contextual information and mine point relations through iterative message passing. 3) Visual Features Augmentation. We introduce a simple yet effective fusion strategy to compensate for sparse LiDAR points with limited semantic cues. Based on these modules, we construct our Graph R-CNN as the second stage, which can be applied to existing one-stage detectors to consistently improve the detection performance. Extensive experiments show that Graph R-CNN outperforms the state-of-the-art 3D detection models by a large margin on both the KITTI and Waymo Open Dataset. And we rank first place on the KITTI BEV car detection leaderboard.
Object pose estimation is a crucial task in computer vision and augmented reality. One of its key challenges is the difficulty of annotation of real training data and the lack of textured CAD models. Therefore, pipelines which do not require CAD models and which can be trained with few labeled images are desirable. We propose a weakly-supervised approach for object pose estimation from RGB-D data using training sets composed of very few labeled images with pose annotations along with weakly-labeled images with ground truth segmentation masks without pose labels. We achieve this by learning to annotate weakly-labeled training data through shape alignment while simultaneously training a pose prediction network. Point cloud alignment is performed using structure and rotation-invariant feature-based losses. We further learn an implicit shape representation, which allows the method to work without the known CAD model and also contributes to pose alignment and pose refinement during training on weakly labeled images. The experimental evaluation shows that our method achieves state-of-the-art results on LineMOD, Occlusion-LineMOD and TLess despite being trained using relative poses and on only a fraction of labeled data used by the other methods. We also achieve comparable results to state-of-the-art RGB-D based pose estimation approaches even when further reducing the amount of unlabeled training data. In addition, our method works even if relative camera poses are given instead of object pose annotations which are typically easier to obtain.
Point clouds and RGB images are two general perceptional sources in autonomous driving. The former can provide accurate localization of objects, and the latter is denser and richer in semantic information. Recently, AutoAlign presents a learnable paradigm in combining these two modalities for 3D object detection. However, it suffers from high computational cost introduced by the global-wise attention. To solve the problem, we propose Cross-Domain DeformCAFA module in this work. It attends to sparse learnable sampling points for cross-modal relational modeling, which enhances the tolerance to calibration error and greatly speeds up the feature aggregation across different modalities. To overcome the complex GT-AUG under multi-modal settings, we design a simple yet effective cross-modal augmentation strategy on convex combination of image patches given their depth information. Moreover, by carrying out a novel image-level dropout training scheme, our model is able to infer in a dynamic manner. To this end, we propose AutoAlignV2, a faster and stronger multi-modal 3D detection framework, built on top of AutoAlign. Extensive experiments on nuScenes benchmark demonstrate the effectiveness and efficiency of AutoAlignV2. Notably, our best model reaches 72.4 NDS on nuScenes test leaderboard, achieving new state-of-the-art results among all published multi-modal 3D object detectors.
Do deep learning models for instance segmentation generalize to novel objects in a systematic way? For classification, such behavior has been questioned. In this study, we aim to understand if certain design decisions such as framework, architecture or pre-training contribute to the semantic understanding of instance segmentation. To answer this question, we consider a special case of robustness and compare pre-trained models on a challenging benchmark for object-centric out-of-distribution texture. We do not introduce another method in this work. Instead, we take a step back and evaluate a broad range of existing literature. This includes Cascade and Mask R-CNN, Swin Transformer, BMask, YOLACT(++), DETR, BCNet, SOTR and SOLOv2. We find that YOLACT++, SOTR and SOLOv2 are significantly more robust to out-of-distribution texture than other frameworks. In addition, we show that deeper and dynamic architectures improve robustness whereas training schedules, data augmentation and pre-training have only a minor impact. In summary we evaluate 68 models on 61 versions of MS COCO for a total of 4148 evaluations.
High-quality structured data with rich annotations are critical components in intelligent vehicle systems dealing with road scenes. However, data curation and annotation require intensive investments and yield low-diversity scenarios. The recently growing interest in synthetic data raises questions about the scope of improvement in such systems and the amount of manual work still required to produce high volumes and variations of simulated data. This work proposes a synthetic data generation pipeline that utilizes existing datasets, like nuScenes, to address the difficulties and domain-gaps present in simulated datasets. We show that using annotations and visual cues from existing datasets, we can facilitate automated multi-modal data generation, mimicking real scene properties with high-fidelity, along with mechanisms to diversify samples in a physically meaningful way. We demonstrate improvements in mIoU metrics by presenting qualitative and quantitative experiments with real and synthetic data for semantic segmentation on the Cityscapes and KITTI-STEP datasets. All relevant code and data is released on github.
We introduce a few-shot localization dataset originating from photographers who authentically were trying to learn about the visual content in the images they took. It includes over 8,000 segmentations of 100 categories in over 4,000 images that were taken by people with visual impairments. Compared to existing few-shot object detection and instance segmentation datasets, our dataset is the first to locate holes in objects (e.g., found in 12.4% of our segmentations), it shows objects that occupy a much larger range of sizes relative to the images, and text is over five times more common in our objects (e.g., found in 24.7% of our segmentations). Analysis of two modern few-shot localization algorithms demonstrates that they generalize poorly to our new dataset. The algorithms commonly struggle to locate objects with holes, very small and very large objects, and objects lacking text. To encourage a larger community to work on these unsolved challenges, we publicly share our annotated few-shot dataset at http://anonymous.
Optical flow estimation in omnidirectional videos faces two significant issues: the lack of benchmark datasets and the challenge of adapting perspective video-based methods to accommodate the omnidirectional nature. This paper proposes the first perceptually natural-synthetic omnidirectional benchmark dataset with a 360° field of view, FLOW360, with 40 different videos and 4,000 video frames. We conduct comprehensive characteristic analysis and comparisons between our dataset and existing optical flow datasets, which manifest perceptual realism, uniqueness, and diversity. To accommodate the omnidirectional nature, we present a novel Siamese representation Learning framework for Omnidirectional Flow (SLOF). We train our network in a contrastive manner with a hybrid loss function that combines contrastive loss and optical flow loss. Extensive experiments verify the proposed framework’s effectiveness and show up to 40% performance improvement over the state-of-the-art approaches. Our FLOW360 dataset and code are available at https://siamlof.github.io/.