Camera and millimeter-wave (MMW) radar fusion is essential for accurate and robust autonomous driving systems. With the advancement of radar technology, next-generation high-resolution automotive radar, i.e., 4-D radar, has emerged. In addition to the target range, azimuth, and Doppler velocity measurements of traditional radar, 4-D radar provides elevation measurement to create a denser “point cloud.” In this study, we propose a camera and 4-D radar fusion network called RCFusion, which achieves multimodal feature fusion under a unified bird’s-eye view (BEV) space to accomplish 3-D object detection tasks. In the camera stream, multiscale feature maps are obtained by the image backbone and feature pyramid network (FPN); they are then converted into orthographic feature maps by an orthographic feature transform (OFT). Next, enhanced and fine-grained image BEV features are obtained via a designed shared attention encoder. Meanwhile, in the 4-D radar stream, a newly designed component named radar PillarNet efficiently encodes the radar features to generate radar pseudo-images, which are fed into the point cloud backbone to create radar BEV features. An interactive attention module (IAM) is proposed for the fusion stage, which outputs a valid fusion of the two-modal BEV features. Finally, a generic detection head predicts the object classes and locations. The proposed RCFusion is validated on the TJ4DRadSet and view-of-delft (VoD) datasets. The experimental results and analysis show that the proposed method can effectively fuse camera and 4-D radar features to achieve robust detection performance.
Although modern object detection and classification models achieve high accuracy, these are typically constrained in advance on a fixed train set and are therefore not flexible to deal with novel, unseen object categories. Moreover, these models most often operate on a single frame, which may yield incorrect classifications in case of ambiguous viewpoints. In this paper, we propose an active inference agent that actively gathers evidence for object classifications, and can learn novel object categories over time. Drawing inspiration from the human brain, we build object-centric generative models composed of two information streams, a what- and a where-stream. The what-stream predicts whether the observed object belongs to a specific category, while the where-stream is responsible for representing the object in its internal 3D reference frame. We show that our agent (i) is able to learn representations for many object categories in an unsupervised way, (ii) achieves state-of-the-art classification accuracies, actively resolving ambiguity when required and (iii) identifies novel object categories. Furthermore, we validate our system in an end-to-end fashion where the agent is able to search for an object at a given pose from a pixel-based rendering. We believe that this is a first step towards building modular, intelligent systems that can be used for a wide range of tasks involving three dimensional objects.
Expensive bounding-box annotations have limited the development of object detection task. Thus, it is necessary to focus on more challenging task of few-shot object detection. It requires the detector to recognize objects of novel classes with only a few training samples. Nowadays, many existing popular methods adopting training way similar to meta-learning have achieved promising performance, such as Meta R-CNN series. However, support data is only used as the class attention to guide the detecting of query images each time. Their relevance to each other remains unexploited. Moreover, a lot of recent works treat the support data and query images as independent branch without considering the relationship between them. To address this issue, we propose a dynamic relevance learning model, which utilizes the relationship between all support images and Region of Interest (RoI) on the query images to construct a dynamic graph convolutional network (GCN). By adjusting the prediction distribution of the base detector using the output of this GCN, the proposed model serves as a hard auxiliary classification task, which guides the detector to improve the class representation implicitly. Comprehensive experiments have been conducted on Pascal VOC and MS-COCO dataset. The proposed model achieves the best overall performance, which shows its effectiveness of learning more generalized features. Our code is available at https://github.com/liuweijie19980216/DRL-for-FSOD.
Previous works for LiDAR-based 3D object detection mainly focus on the single-frame paradigm. In this paper, we propose to detect 3D objects by exploiting temporal information in multiple frames, i.e., the point cloud videos. We empirically categorize the temporal information into short-term and long-term patterns. To encode the short-term data, we present a Grid Message Passing Network (GMPNet), which considers each grid (i.e., the grouped points) as a node and constructs a k-NN graph with the neighbor grids. To update features for a grid, GMPNet iteratively collects information from its neighbors, thus mining the motion cues in grids from nearby frames. To further aggregate the long-term frames, we propose an Attentive Spatiotemporal Transformer GRU (AST-GRU), which contains a Spatial Transformer Attention (STA) module and a Temporal Transformer Attention (TTA) module. STA and TTA enhance the vanilla GRU to focus on small objects and better align the moving objects. Our overall framework supports both online and offline video object detection in point clouds. We implement our algorithm based on prevalent anchor-based and anchor-free detectors. The evaluation results on the challenging nuScenes benchmark show the superior performance of our method, achieving the 1st on the leaderboard without any bells and whistles, by the time the paper is submitted.
Apples growing in natural environments often face severe visual obstructions from leaves and branches. This significantly increases the risk of false detections in object detection tasks, thereby escalating the challenge. Addressing this issue, we introduce a technique called "Occlusion-Enhanced Distillation" (OED). This approach utilizes occlusion information to regularize the learning of semantically aligned features on occluded datasets and employs Exponential Moving Average (EMA) to enhance training stability. Specifically, we first design an occlusion-enhanced dataset that integrates Grounding DINO and SAM methods to extract occluding elements such as leaves and branches from each sample, creating occlusion examples that reflect the natural growth state of fruits. Additionally, we propose a multi-scale knowledge distillation strategy, where the student network uses images with increased occlusions as inputs, while the teacher network employs images without natural occlusions. Through this setup, the strategy guides the student network to learn from the teacher across scales of semantic and local features alignment, effectively narrowing the feature distance between occluded and non-occluded targets and enhancing the robustness of object detection. Lastly, to improve the stability of the student network, we introduce the EMA strategy, which aids the student network in learning more generalized feature expressions that are less affected by the noise of individual image occlusions. Our method significantly outperforms current state-of-the-art techniques through extensive comparative experiments.
Active perception describes a broad class of techniques that couple planning and perception systems to move the robot in a way to give the robot more information about the environment. In most robotic systems, perception is typically independent of motion planning. For example, traditional object detection is passive: it operates only on the images it receives. However, we have a chance to improve the results if we allow planning to consume detection signals and move the robot to collect views that maximize the quality of the results. In this paper, we use reinforcement learning (RL) methods to control the robot in order to obtain images that maximize the detection quality. Specifically, we propose using a Decision Transformer with online fine-tuning, which first optimizes the policy with a pre-collected expert dataset and then improves the learned policy by exploring better solutions in the environment. We evaluate the performance of proposed method on an interactive dataset collected from an indoor scenario simulator. Experimental results demonstrate that our method outperforms all baselines, including expert policy and pure offline RL methods. We also provide exhaustive analyses of the reward distribution and observation space.
Multi-task networks can potentially improve performance and computational efficiency compared to single-task networks, facilitating online deployment. However, current multi-task architectures in point cloud perception combine multiple task-specific point cloud representations, each requiring a separate feature encoder and making the network structures bulky and slow. We propose PAttFormer, an efficient multi-task architecture for joint semantic segmentation and object detection in point clouds that only relies on a point-based representation. The network builds on transformer-based feature encoders using neighborhood attention and grid-pooling and a query-based detection decoder using a novel 3D deformable-attention detection head design. Unlike other LiDAR-based multi-task architectures, our proposed PAttFormer does not require separate feature encoders for multiple task-specific point cloud representations, resulting in a network that is 3x smaller and 1.4x faster while achieving competitive performance on the nuScenes and KITTI benchmarks for autonomous driving perception. Our extensive evaluations show substantial gains from multi-task learning, improving LiDAR semantic segmentation by +1.7% in mIou and 3D object detection by +1.7% in mAP on the nuScenes benchmark compared to the single-task models.
Image-based Pose-Agnostic 3D Anomaly Detection is an important task that has emerged in industrial quality control. This task seeks to find anomalies from query images of a tested object given a set of reference images of an anomaly-free object. The challenge is that the query views (a.k.a poses) are unknown and can be different from the reference views. Currently, new methods such as OmniposeAD and SplatPose have emerged to bridge the gap by synthesizing pseudo reference images at the query views for pixel-to-pixel comparison. However, none of these methods can infer in real-time, which is critical in industrial quality control for massive production. For this reason, we propose SplatPose+, which employs a hybrid representation consisting of a Structure from Motion (SfM) model for localization and a 3D Gaussian Splatting (3DGS) model for Novel View Synthesis. Although our proposed pipeline requires the computation of an additional SfM model, it offers real-time inference speeds and faster training compared to SplatPose. Quality-wise, we achieved a new SOTA on the Pose-agnostic Anomaly Detection benchmark with the Multi-Pose Anomaly Detection (MAD-SIM) dataset.
We introduced a high-resolution equirectangular panorama (360-degree, virtual reality) dataset for object detection and propose a multi-projection variant of YOLO detector. The main challenge with equirectangular panorama image are i) the lack of annotated training data, ii) high-resolution imagery and iii) severe geometric distortions of objects near the panorama projection poles. In this work, we solve the challenges by i) using training examples available in the "conventional datasets" (ImageNet and COCO), ii) employing only low-resolution images that require only moderate GPU computing power and memory, and iii) our multi-projection YOLO handles projection distortions by making multiple stereographic sub-projections. In our experiments, YOLO outperforms the other state-of-art detector, Faster RCNN and our multi-projection YOLO achieves the best accuracy with low-resolution input.
This report details our solution to the Google AI Open Images Challenge 2019 Object Detection Track. Based on our detailed analysis on the Open Images dataset, it is found that there are four typical features: large-scale, hierarchical tag system, severe annotation incompleteness and data imbalance. Considering these characteristics, many strategies are employed, including larger backbone, distributed softmax loss, class-aware sampling, expert model, and heavier classifier. In virtue of these effective strategies, our best single model could achieve a mAP of 61.90. After ensemble, the final mAP is boosted to 67.17 in the public leaderboard and 64.21 in the private leaderboard, which earns 3rd place in the Open Images Challenge 2019.
The rotation prediction (Rotation) is a simple pretext-task for self-supervised learning (SSL), where models learn useful representations for target vision tasks by solving pretext-tasks. Although Rotation captures information of object shapes, it hardly captures information of textures. To tackle this problem, we introduce a novel pretext-task called image enhanced rotation prediction (IE-Rot) for SSL. IE-Rot simultaneously solves Rotation and another pretext-task based on image enhancement (e.g., sharpening and solarizing) while maintaining simplicity. Through the simultaneous prediction of rotation and image enhancement, models learn representations to capture the information of not only object shapes but also textures. Our experimental results show that IE-Rot models outperform Rotation on various standard benchmarks including ImageNet classification, PASCAL-VOC detection, and COCO detection/segmentation.
Wood species identification plays a crucial role in various industries, from ensuring the legality of timber products to advancing ecological conservation efforts. This paper introduces WoodYOLO, a novel object detection algorithm specifically designed for microscopic wood fiber analysis. Our approach adapts the YOLO architecture to address the challenges posed by large, high-resolution microscopy images and the need for high recall in localization of the cell type of interest (vessel elements). Our results show that WoodYOLO significantly outperforms state-of-the-art models, achieving performance gains of 12.9% and 6.5% in F2 score over YOLOv10 and YOLOv7, respectively. This improvement in automated wood cell type localization capabilities contributes to enhancing regulatory compliance, supporting sustainable forestry practices, and promoting biodiversity conservation efforts globally.
Vision-based object detectors are a crucial basis for robotics applications as they provide valuable information about object localisation in the environment. These need to ensure high reliability in different lighting conditions, occlusions, and visual artifacts, all while running in real-time. Collecting and annotating real-world data for these networks is prohibitively time consuming and costly, especially for custom assets, such as industrial objects, making it untenable for generalization to in-the-wild scenarios. To this end, we present Synthetica, a method for large-scale synthetic data generation for training robust state estimators. This paper focuses on the task of object detection, an important problem which can serve as the front-end for most state estimation problems, such as pose estimation. Leveraging data from a photorealistic ray-tracing renderer, we scale up data generation, generating 2.7 million images, to train highly accurate real-time detection transformers. We present a collection of rendering randomization and training-time data augmentation techniques conducive to robust sim-to-real performance for vision tasks. We demonstrate state-of-the-art performance on the task of object detection while having detectors that run at 50-100Hz which is 9 times faster than the prior SOTA. We further demonstrate the usefulness of our training methodology for robotics applications by showcasing a pipeline for use in the real world with custom objects for which there do not exist prior datasets. Our work highlights the importance of scaling synthetic data generation for robust sim-to-real transfer while achieving the fastest real-time inference speeds. Videos and supplementary information can be found at this URL: https://sites.google.com/view/synthetica-vision.
There is a growing demand for accurate high-resolution land cover maps in many fields, e.g., in land-use planning and biodiversity conservation. Developing such maps has been performed using Object-Based Image Analysis (OBIA) methods, which usually reach good accuracies, but require a high human supervision and the best configuration for one image can hardly be extrapolated to a different image. Recently, the deep learning Convolutional Neural Networks (CNNs) have shown outstanding results in object recognition in the field of computer vision. However, they have not been fully explored yet in land cover mapping for detecting species of high biodiversity conservation interest. This paper analyzes the potential of CNNs-based methods for plant species detection using free high-resolution Google Earth T M images and provides an objective comparison with the state-of-the-art OBIA-methods. We consider as case study the detection of Ziziphus lotus shrubs, which are protected as a priority habitat under the European Union Habitats Directive. According to our results, compared to OBIA-based methods, the proposed CNN-based detection model, in combination with data-augmentation, transfer learning and pre-processing, achieves higher performance with less human intervention and the knowledge it acquires in the first image can be transferred to other images, which makes the detection process very fast. The provided methodology can be systematically reproduced for other species detection.
We present 3D-MPA, a method for instance segmentation on 3D point clouds. Given an input point cloud, we propose an object-centric approach where each point votes for its object center. We sample object proposals from the predicted object centers. Then, we learn proposal features from grouped point features that voted for the same object center. A graph convolutional network introduces inter-proposal relations, providing higher-level feature learning in addition to the lower-level point features. Each proposal comprises a semantic label, a set of associated points over which we define a foreground-background mask, an objectness score and aggregation features. Previous works usually perform non-maximum-suppression (NMS) over proposals to obtain the final object detections or semantic instances. However, NMS can discard potentially correct predictions. Instead, our approach keeps all proposals and groups them together based on the learned aggregation features. We show that grouping proposals improves over NMS and outperforms previous state-of-the-art methods on the tasks of 3D object detection and semantic instance segmentation on the ScanNetV2 benchmark and the S3DIS dataset.
Massive semantically labeled datasets are readily available for 2D images, however, are much harder to achieve for 3D scenes. Objects in 3D repositories like ShapeNet are labeled, but regrettably only in isolation, so without context. 3D scenes can be acquired by range scanners on city-level scale, but much fewer with semantic labels. Addressing this disparity, we introduce a new optimization procedure, which allows training for 3D detection with raw 3D scans while using as little as 5% of the object labels and still achieve comparable performance. Our optimization uses two networks. A scene network maps an entire 3D scene to a set of 3D object centers. As we assume the scene not to be labeled by centers, no classic loss, such as Chamfer can be used to train it. Instead, we use another network to emulate the loss. This loss network is trained on a small labeled subset and maps a non centered 3D object in the presence of distractions to its own center. This function is very similar - and hence can be used instead of - the gradient the supervised loss would provide. Our evaluation documents competitive fidelity at a much lower level of supervision, respectively higher quality at comparable supervision. Supplementary material can be found at: https://dgriffiths3.github.io.
Object detection and global localization play a crucial role in robotics, spanning across a great spectrum of applications from autonomous cars to multi-layered 3D Scene Graphs for semantic scene understanding. This article proposes BOX3D, a novel multi-modal and lightweight scheme for localizing objects of interest by fusing the information from RGB camera and 3D LiDAR. BOX3D is structured around a three-layered architecture, building up from the local perception of the incoming sequential sensor data to the global perception refinement that covers for outliers and the general consistency of each object's observation. More specifically, the first layer handles the low-level fusion of camera and LiDAR data for initial 3D bounding box extraction. The second layer converts each LiDAR's scan 3D bounding boxes to the world coordinate frame and applies a spatial pairing and merging mechanism to maintain the uniqueness of objects observed from different viewpoints. Finally, BOX3D integrates the third layer that supervises the consistency of the results on the global map iteratively, using a point-to-voxel comparison for identifying all points in the global map that belong to the object. Benchmarking results of the proposed novel architecture are showcased in multiple experimental trials on public state-of-the-art large-scale dataset of urban environments.
객체 검출, 분할, 장면 그래프 생성과 같은 구조화된 시각 인식 작업을 위한 기존 지도 학습 방법은 종종 결정론적이고 고정된 출력을 생성하여, 복잡한 시각 장면에 내재된 불확실성을 포착하는 능력을 제한한다. 그 결과, 이러한 점 추정은 부분적으로 가려진 객체의 크기 모호성, 정확한 분할 경계의 국소적 모호성 등 자연적 모호성뿐만 아니라 노이즈와 학습 데이터의 희소성에서 비롯되는, 이러한 문제에 내재된 예측 불확실성(또는 다중 모달성)을 포착하지 못한다. 이 한계를 해결하기 위해, 우리는 주어진 입력 이미지에 대해 구조화된 출력에 대한 분포를 학습하는 간단하고 새로운 프레임워크인 Modular Diffusion Models(MDM)를 제시한다. MDM은 확산 과정을 각각 객체 범주, 공간적 위치, 객체 간 관계 등 구조화된 정보 공간의 서로 다른 측면을 포착하는 데 집중하는, 서로 구별되는 과제별 모듈로 분해한다. 이 모듈형 설계는 각 구성 요소를 독립적으로 학습할 수 있게 하며, 추가 학습 없이 추론 시 원활하게 통합할 수 있다. 또한 MDM의 모듈성은 확산 과정이 많은 구조화 학습 작업에서 흔한 이질적 출력 공간(예: 연속적인 바운딩 박스와 이산적 클래스 레이블) 위에서 쉽게 작동할 수 있게 한다. 객체 검출, 인스턴스 분할, 장면 그래프 생성이라는 세 가지 서로 다른 구조화 작업에 대한 실험 결과는 우리가 제안한 프레임워크의 이점을 보여준다.
In recent years, many deep learning models have been adopted in autonomous driving. At the same time, these models introduce new vulnerabilities that may compromise the safety of autonomous vehicles. Specifically, recent studies have demonstrated that adversarial attacks can cause a significant decline in detection precision of deep learning-based 3D object detection models. Although driving safety is the ultimate concern for autonomous driving, there is no comprehensive study on the linkage between the performance of deep learning models and the driving safety of autonomous vehicles under adversarial attacks. In this paper, we investigate the impact of two primary types of adversarial attacks, perturbation attacks and patch attacks, on the driving safety of vision-based autonomous vehicles rather than the detection precision of deep learning models. In particular, we consider two state-of-the-art models in vision-based 3D object detection, Stereo R-CNN and DSGN. To evaluate driving safety, we propose an end-to-end evaluation framework with a set of driving safety performance metrics. By analyzing the results of our extensive evaluation experiments, we find that (1) the attack's impact on the driving safety of autonomous vehicles and the attack's impact on the precision of 3D object detectors are decoupled, and (2) the DSGN model demonstrates stronger robustness to adversarial attacks than the Stereo R-CNN model. In addition, we further investigate the causes behind the two findings with an ablation study. The findings of this paper provide a new perspective to evaluate adversarial attacks and guide the selection of deep learning models in autonomous driving.
In the photolithographic process vital to semiconductor manufacturing, various types of defects appear during EUV pattering. Due to ever-shrinking pattern size, these defects are extremely small and cause false or missed detection during inspection. Specifically, the lack of defect-annotated quality data with good representation of smaller defects has prohibited deployment of deep learning based defect detection models in fabrication lines. To resolve the problem of data unavailability, we artificially generate scanning electron microscopy (SEM) images of line patterns with known distribution of defects and autonomously annotate them. We then employ state-of-the-art object detection models to investigate defect detection performance as a function of defect size, much smaller than the pitch width. We find that the real-time object detector YOLOv8 has the best mean average precision of 96% as compared to EfficientNet, 83%, and SSD, 77%, with the ability to detect smaller defects. We report the smallest defect size that can be detected reliably. When tested on real SEM data, the YOLOv8 model correctly detected 84.6% of Bridge defects and 78.3% of Break defects across all relevant instances. These promising results suggest that synthetic data can be used as an alternative to real-world data in order to develop robust machine-learning models.