Artifacts in imagery captured by remote sensing, such as clouds, snow, and shadows, present challenges for various tasks, including semantic segmentation and object detection. A primary challenge in developing algorithms for identifying such artifacts is the cost of collecting annotated training data. In this work, we explore how recent advances in multi-image fusion can be leveraged to bootstrap single image cloud detection. We demonstrate that a network optimized to estimate image quality also implicitly learns to detect clouds. To support the training and evaluation of our approach, we collect a large dataset of Sentinel-2 images along with a per-pixel semantic labelling for land cover. Through various experiments, we demonstrate that our method reduces the need for annotated training data and improves cloud detection performance.
The thesis contributes in several important ways to the research area of 3D object category learning and recognition. To cope with the mentioned limitations, we look at human cognition, in particular at the fact that human beings learn to recognize object categories ceaselessly over time. This ability to refine knowledge from the set of accumulated experiences facilitates the adaptation to new environments. Inspired by this capability, we seek to create a cognitive object perception and perceptual learning architecture that can learn 3D object categories in an open-ended fashion. In this context, ``open-ended'' implies that the set of categories to be learned is not known in advance, and the training instances are extracted from actual experiences of a robot, and thus become gradually available, rather than being available since the beginning of the learning process. In particular, this architecture provides perception capabilities that will allow robots to incrementally learn object categories from the set of accumulated experiences and reason about how to perform complex tasks. This framework integrates detection, tracking, teaching, learning, and recognition of objects. An extensive set of systematic experiments, in multiple experimental settings, was carried out to thoroughly evaluate the described learning approaches. Experimental results show that the proposed system is able to interact with human users, learn new object categories over time, as well as perform complex tasks. The contributions presented in this thesis have been fully implemented and evaluated on different standard object and scene datasets and empirically evaluated on different robotic platforms.
Many computer vision tasks address the problem of scene understanding and are naturally interrelated e.g. object classification, detection, scene segmentation, depth estimation, etc. We show that we can leverage the inherent relationships among collections of tasks, as they are trained jointly, supervising each other through their known relationships via consistency losses. Furthermore, explicitly utilizing the relationships between tasks allows improving their performance while dramatically reducing the need for labeled data, and allows training with additional unsupervised or simulated data. We demonstrate a distributed joint training algorithm with task-level parallelism, which affords a high degree of asynchronicity and robustness. This allows learning across multiple tasks, or with large amounts of input data, at scale. We demonstrate our framework on subsets of the following collection of tasks: depth and normal prediction, semantic segmentation, 3D motion and ego-motion estimation, and object tracking and 3D detection in point clouds. We observe improved performance across these tasks, especially in the low-label regime.
This paper introduces a concept of layer aggregation to describe how information from previous layers can be reused to better extract features at the current layer. While DenseNet is a typical example of the layer aggregation mechanism, its redundancy has been commonly criticized in the literature. This motivates us to propose a very light-weighted module, called recurrent layer aggregation (RLA), by making use of the sequential structure of layers in a deep CNN. Our RLA module is compatible with many mainstream deep CNNs, including ResNets, Xception and MobileNetV2, and its effectiveness is verified by our extensive experiments on image classification, object detection and instance segmentation tasks. Specifically, improvements can be uniformly observed on CIFAR, ImageNet and MS COCO datasets, and the corresponding RLA-Nets can surprisingly boost the performances by 2-3% on the object detection task. This evidences the power of our RLA module in helping main CNNs better learn structural information in images.
Regression loss design is an essential topic for oriented object detection. Due to the periodicity of the angle and the ambiguity of width and height definition, traditional L1-distance loss and its variants have been suffered from the metric discontinuity and the square-like problem. As a solution, the distribution based methods show significant advantages by representing oriented boxes as distributions. Differing from exploited the Gaussian distribution to get analytical form of distance measure, we propose a novel oriented regression loss, Wasserstein Distance(EWD) loss, to alleviate the square-like problem. Specifically, for the oriented box(OBox) representation, we choose a specially-designed distribution whose probability density function is only nonzero over the edges. On this basis, we develop Wasserstein distance as the measure. Besides, based on the edge representation of OBox, the EWD loss can be generalized to quadrilateral and polynomial regression scenarios. Experiments on multiple popular datasets and different detectors show the effectiveness of the proposed method.
PETR-based methods have dominated benchmarks in 3D perception and are increasingly becoming a key component in modern autonomous driving systems. However, their quantization performance significantly degrades when INT8 inference is required, with a degradation of 58.2% in mAP and 36.9% in NDS on the NuScenes dataset. To address this issue, we propose a quantization-aware position embedding transformation for multi-view 3D object detection, termed Q-PETR. Q-PETR offers a quantizationfriendly and deployment-friendly architecture while preserving the original performance of PETR. It substantially narrows the accuracy gap between INT8 and FP32 inference for PETR-series methods. Without bells and whistles, our approach reduces the mAP and NDS drop to within 1% under standard 8-bit per-tensor post-training quantization. Furthermore, our method exceeds the performance of the original PETR in terms of floating-point precision. Extensive experiments across a variety of PETR-series models demonstrate its broad generalization.
Most applications in autonomous navigation using mounted cameras rely on the construction and processing of geometric 3D point clouds, which is an expensive process. However, there is another simpler way to make a space navigable quickly: to use semantic information (e.g., traffic signs) to guide the agent. However, detecting and acting on semantic information involves Computer Vision~(CV) algorithms such as object detection, which themselves are demanding for agents such as aerial drones with limited onboard resources. To solve this problem, we introduce a novel Markov Decision Process~(MDP) framework to reduce the workload of these CV approaches. We apply our proposed framework to both feature-based and neural-network-based object-detection tasks, using open-loop and closed-loop simulations as well as hardware-in-the-loop emulations. These holistic tests show significant benefits in energy consumption and speed with only a limited loss in accuracy compared to models based on static features and neural networks.
적외선 영상 초해상화(IISR)는 객체 검출(object detection) 및 시맨틱 세그멘테이션과 같은 다운스트림 태스크에 중요합니다. 기존 IISR 방법들은 인위적인 텍스처, 과도하게 날카로워진 경계선, 위조된 고주파 세부 정보를 생성하여 실제 열적 구조와 의미 정보를 왜곡하는 경우가 많습니다. 이러한 문제를 해결하기 위해, 신뢰할 수 있는 기계 인식을 위한 충실한 적외선 초해상화 프레임워크인 FaithIR을 제안합니다. FaithIR은 전역적인 열 및 구조 정보를 포착하는 패치 수준 컨디셔닝 브랜치와 구조적 가이드 하에 조밀한 국소 재구성을 수행하는 픽셀 수준 복원 브랜치로 구성됩니다. 전체 복원 과정은 적외선 고유의 구조와 태스크 관련 정보를 보존하기 위해 픽셀 도메인에서 직접 수행됩니다. FLIR-IISR, M3FD, FMB 데이터셋에서의 광범위한 실험을 통해 우수한 재구성 충실도, 데이터셋 간 일반화 성능, 객체 검출 및 시맨틱 세그멘테이션에서의 뛰어난 성능을 입증합니다. 이러한 결과는 신뢰할 수 있는 기계 인식을 위해 단순히 지각적 선명도만을 추구하는 것보다 충실한 적외선 구조를 보존하는 것이 더 중요하다는 점을 보여줍니다.
Top-down saliency models produce a probability map that peaks at target locations specified by a task/goal such as object detection. They are usually trained in a fully supervised setting involving pixel-level annotations of objects. We propose a weakly supervised top-down saliency framework using only binary labels that indicate the presence/absence of an object in an image. First, the probabilistic contribution of each image region to the confidence of a CNN-based image classifier is computed through a backtracking strategy to produce top-down saliency. From a set of saliency maps of an image produced by fast bottom-up saliency approaches, we select the best saliency map suitable for the top-down task. The selected bottom-up saliency map is combined with the top-down saliency map. Features having high combined saliency are used to train a linear SVM classifier to estimate feature saliency. This is integrated with combined saliency and further refined through a multi-scale superpixel-averaging of saliency map. We evaluate the performance of the proposed weakly supervised topdown saliency and achieve comparable performance with fully supervised approaches. Experiments are carried out on seven challenging datasets and quantitative results are compared with 40 closely related approaches across 4 different applications.
We present a conceptually simple, flexible, and universal visual perception head for variant visual tasks, e.g., classification, object detection, instance segmentation and pose estimation, and different frameworks, such as one-stage or two-stage pipelines. Our approach effectively identifies an object in an image while simultaneously generating a high-quality bounding box or contour-based segmentation mask or set of keypoints. The method, called UniHead, views different visual perception tasks as the dispersible points learning via the transformer encoder architecture. Given a fixed spatial coordinate, UniHead adaptively scatters it to different spatial points and reasons about their relations by transformer encoder. It directly outputs the final set of predictions in the form of multiple points, allowing us to perform different visual tasks in different frameworks with the same head design. We show extensive evaluations on ImageNet classification and all three tracks of the COCO suite of challenges, including object detection, instance segmentation and pose estimation. Without bells and whistles, UniHead can unify these visual tasks via a single visual head design and achieve comparable performance compared to expert models developed for each task.We hope our simple and universal UniHead will serve as a solid baseline and help promote universal visual perception research. Code and models are available at https://github.com/Sense-X/UniHead.
Knowledge Distillation (KD) for object detection aims to train a compact detector by transferring knowledge from a teacher model. Since the teacher model perceives data in a way different from humans, existing KD methods only distill knowledge that is consistent with labels annotated by human expert while neglecting knowledge that is not consistent with human perception, which results in insufficient distillation and sub-optimal performance. In this paper, we propose inconsistent knowledge distillation (IKD), which aims to distill knowledge inherent in the teacher model's counter-intuitive perceptions. We start by considering the teacher model's counter-intuitive perceptions of frequency and non-robust features. Unlike previous works that exploit fine-grained features or introduce additional regularizations, we extract inconsistent knowledge by providing diverse input using data augmentation. Specifically, we propose a sample-specific data augmentation to transfer the teacher model's ability in capturing distinct frequency components and suggest an adversarial feature augmentation to extract the teacher model's perceptions of non-robust features in the data. Extensive experiments demonstrate the effectiveness of our method which outperforms state-of-the-art KD baselines on one-stage, two-stage and anchor-free object detectors (at most +1.0 mAP). Our codes will be made available at \url{https://github.com/JWLiang007/IKD.git}.
We tackle the problem of novel class discovery and localization (NCDL). In this setting, we assume a source dataset with supervision for only some object classes. Instances of other classes need to be discovered, classified, and localized automatically based on visual similarity without any human supervision. To tackle NCDL, we propose a two-stage object detection network Region-based NCDL (RNCDL) that uses a region proposal network to localize regions of interest (RoIs). We then train our network to learn to classify each RoI, either as one of the known classes, seen in the source dataset, or one of the novel classes, with a long-tail distribution constraint on the class assignments, reflecting the natural frequency of classes in the real world. By training our detection network with this objective in an end-to-end manner, it learns to classify all region proposals for a large variety of classes, including those not part of the labeled object class vocabulary. Our experiments conducted using COCO and LVIS datasets reveal that our method is significantly more effective than multi-stage pipelines that rely on traditional clustering algorithms. Furthermore, we demonstrate the generality of our approach by applying our method to a large-scale Visual Genome dataset, where our network successfully learns to detect various semantic classes without direct supervision.
건설 현장의 비전 기반 모니터링은 밤, 새벽과 해 질 무렵, 터널 내부, 투광 조명 아래에서 실패할 수 있다. 일반적인 RGB 카메라에 노출 부족이나 과다 노출이 생기기 때문이다. 이벤트 카메라는 픽셀마다 밝기 변화를 마이크로초 단위의 지연 시간과 120dB를 넘는 동적 범위로 기록한다. 그러나 실제 데이터에서 현장에 필요한 객체 검출에 얼마나 유용한지는 정량적으로 평가되지 않았다. 본 연구는 픽셀 단위로 정렬된 PEOD 데이터셋을 사용해 최근 이벤트 기반 검출기 세 종(SMamba, SATE, EvRT-DETR)을 동일한 데이터로 학습한 RGB 검출기와 비교했다. 테스트 세트의 70%인 저조도 프레임에서 최고 성능의 이벤트 검출기는 27.2AP를, RGB 검출기는 21.3AP를 기록했다. 별도 학습 없이 두 센서의 결과를 후반에 융합한 방식은 모든 조명 조건에서 가장 우수했으며 전체 AP는 33.3이었다. 객체 53,091개를 개별 분석한 결과, 밝기는 이벤트 검출기에 거의 영향을 주지 않았지만 빠른 움직임은 영향을 줬다. 프레임당 20픽셀을 넘게 움직이는 중간 크기 객체에서 RGB는 11~19%, EvRT-DETR은 39~60%를 찾았다. 활성화 패치 분석에서 이벤트 정보 손실의 영향은 해상도가 16분의 1인 특징에서 확인됐고, 작은 객체는 가장 세밀한 스케일에 의존했다. 신호 이벤트가 적은 상태에서 잡음이 더해지면 이벤트 검출이 실패했다. 모델에 독립적인 입력 가지치기 실험에서는 입력의 60%만으로 정확도의 약 95%를 유지했으나 작은 객체의 재현율이 절반으로 줄었다. 렌더링한 건설 현장을 제로샷으로 평가했을 때, 투광 조명을 어둡게 하자 RGB의 성능은 0에 가까워졌지만 이벤트 검출기는 AP50의 대부분을 유지했다. 마지막으로 건설 현장의 카메라 배치 지침을 제시한다.
For years, the YOLO series has been the de facto industry-level standard for efficient object detection. The YOLO community has prospered overwhelmingly to enrich its use in a multitude of hardware platforms and abundant scenarios. In this technical report, we strive to push its limits to the next level, stepping forward with an unwavering mindset for industry application. Considering the diverse requirements for speed and accuracy in the real environment, we extensively examine the up-to-date object detection advancements either from industry or academia. Specifically, we heavily assimilate ideas from recent network design, training strategies, testing techniques, quantization, and optimization methods. On top of this, we integrate our thoughts and practice to build a suite of deployment-ready networks at various scales to accommodate diversified use cases. With the generous permission of YOLO authors, we name it YOLOv6. We also express our warm welcome to users and contributors for further enhancement. For a glimpse of performance, our YOLOv6-N hits 35.9% AP on the COCO dataset at a throughput of 1234 FPS on an NVIDIA Tesla T4 GPU. YOLOv6-S strikes 43.5% AP at 495 FPS, outperforming other mainstream detectors at the same scale~(YOLOv5-S, YOLOX-S, and PPYOLOE-S). Our quantized version of YOLOv6-S even brings a new state-of-the-art 43.3% AP at 869 FPS. Furthermore, YOLOv6-M/L also achieves better accuracy performance (i.e., 49.5%/52.3%) than other detectors with a similar inference speed. We carefully conducted experiments to validate the effectiveness of each component. Our code is made available at https://github.com/meituan/YOLOv6.
Generating 3D point cloud (PC) data from noisy sonar measurements is a problem that has potential applications for bathymetry mapping, artificial object inspection, mapping of aquatic plants and fauna as well as underwater navigation and localization of vehicles such as submarines. Side-scan sonar sensors are available in inexpensive cost ranges, especially in fish-finders, where the transducers are usually mounted to the bottom of a boat and can approach shallower depths than the ones attached to an Uncrewed Underwater Vehicle (UUV) can. However, extracting 3D information from side-scan sonar imagery is a difficult task because of its low signal-to-noise ratio and missing angle and depth information in the imagery. Since most algorithms that generate a 3D point cloud from side-scan sonar imagery use Shape from Shading (SFS) techniques, extracting 3D information is especially difficult when the seafloor is smooth, is slowly changing in depth, or does not have identifiable objects that make acoustic shadows. This paper introduces an efficient algorithm that generates a sparse 3D point cloud from side-scan sonar images. This computation is done in a computationally efficient manner by leveraging the geometry of the first sonar return combined with known positions provided by GPS and down-scan sonar depth measurement at each data point. Additionally, this paper implements another algorithm that uses a Convolutional Neural Network (CNN) using transfer learning to perform object detection on side-scan sonar images collected in real life and generated with a simulation. The algorithm was tested on both real and synthetic images to show reasonably accurate anomaly detection and classification.
Open-vocabulary object perception has become an important topic in artificial intelligence, which aims to identify objects with novel classes that have not been seen during training. Under this setting, open-vocabulary object detection (OVD) in a single image has been studied in many literature. However, open-vocabulary object tracking (OVT) from a video has been studied less, and one reason is the shortage of benchmarks. In this work, we have built a new large-scale benchmark for open-vocabulary multi-object tracking namely OVT-B. OVT-B contains 1,048 categories of objects and 1,973 videos with 637,608 bounding box annotations, which is much larger than the sole open-vocabulary tracking dataset, i.e., OVTAO-val dataset (200+ categories, 900+ videos). The proposed OVT-B can be used as a new benchmark to pave the way for OVT research. We also develop a simple yet effective baseline method for OVT. It integrates the motion features for object tracking, which is an important feature for MOT but is ignored in previous OVT methods. Experimental results have verified the usefulness of the proposed benchmark and the effectiveness of our method. We have released the benchmark to the public at https://github.com/Coo1Sea/OVT-B-Dataset.
Semantic segmentation and object detection research have recently achieved rapid progress. However, the former task has no notion of different instances of the same object, and the latter operates at a coarse, bounding-box level. We propose an Instance Segmentation system that produces a segmentation map where each pixel is assigned an object class and instance identity label. Most approaches adapt object detectors to produce segments instead of boxes. In contrast, our method is based on an initial semantic segmentation module, which feeds into an instance subnetwork. This subnetwork uses the initial category-level segmentation, along with cues from the output of an object detector, within an end-to-end CRF to predict instances. This part of our model is dynamically instantiated to produce a variable number of instances per image. Our end-to-end approach requires no post-processing and considers the image holistically, instead of processing independent proposals. Therefore, unlike some related work, a pixel cannot belong to multiple instances. Furthermore, far more precise segmentations are achieved, as shown by our state-of-the-art results (particularly at high IoU thresholds) on the Pascal VOC and Cityscapes datasets.
Most self-supervised learning (SSL) methods often work on curated datasets where the object-centric assumption holds. This assumption breaks down in uncurated images. Existing scene image SSL methods try to find the two views from original scene images that are well matched or dense, which is both complex and computationally heavy. This paper proposes a conceptually different pipeline: first find regions that are coarse objects (with adequate objectness), crop them out as pseudo object-centric images, then any SSL method can be directly applied as in a real object-centric dataset. That is, coarse crops benefits scene images SSL. A novel cropping strategy that produces coarse object box is proposed. The new pipeline and cropping strategy successfully learn quality features from uncurated datasets without ImageNet. Experiments show that our pipeline outperforms existing SSL methods (MoCo-v2, DenseCL and MAE) on classification, detection and segmentation tasks. We further conduct extensively ablations to verify that: 1) the pipeline do not rely on pretrained models; 2) the cropping strategy is better than existing object discovery methods; 3) our method is not sensitive to hyperparameters and data augmentations.
Camera and radar sensors have significant advantages in cost, reliability, and maintenance compared to LiDAR. Existing fusion methods often fuse the outputs of single modalities at the result-level, called the late fusion strategy. This can benefit from using off-the-shelf single sensor detection algorithms, but late fusion cannot fully exploit the complementary properties of sensors, thus having limited performance despite the huge potential of camera-radar fusion. Here we propose a novel proposal-level early fusion approach that effectively exploits both spatial and contextual properties of camera and radar for 3D object detection. Our fusion framework first associates image proposal with radar points in the polar coordinate system to efficiently handle the discrepancy between the coordinate system and spatial properties. Using this as a first stage, following consecutive cross-attention based feature fusion layers adaptively exchange spatio-contextual information between camera and radar, leading to a robust and attentive fusion. Our camera-radar fusion approach achieves the state-of-the-art 41.1% mAP and 52.3% NDS on the nuScenes test set, which is 8.7 and 10.8 points higher than the camera-only baseline, as well as yielding competitive performance on the LiDAR method.
Monocular 3D object detection has become a mainstream approach in automatic driving for its easy application. A prominent advantage is that it does not need LiDAR point clouds during the inference. However, most current methods still rely on 3D point cloud data for labeling the ground truths used in the training phase. This inconsistency between the training and inference makes it hard to utilize the large-scale feedback data and increases the data collection expenses. To bridge this gap, we propose a new weakly supervised monocular 3D objection detection method, which can train the model with only 2D labels marked on images. To be specific, we explore three types of consistency in this task, i.e. the projection, multi-view and direction consistency, and design a weakly-supervised architecture based on these consistencies. Moreover, we propose a new 2D direction labeling method in this task to guide the model for accurate rotation direction prediction. Experiments show that our weakly-supervised method achieves comparable performance with some fully supervised methods. When used as a pre-training method, our model can significantly outperform the corresponding fully-supervised baseline with only 1/3 3D labels. https://github.com/weakmono3d/weakmono3d