Most existing object instance detection and segmentation models only work well on fairly balanced benchmarks where per-category training sample numbers are comparable, such as COCO. They tend to suffer performance drop on realistic datasets that are usually long-tailed. This work aims to study and address such open challenges. Specifically, we systematically investigate performance drop of the state-of-the-art two-stage instance segmentation model Mask R-CNN on the recent long-tail LVIS dataset, and unveil that a major cause is the inaccurate classification of object proposals. Based on such an observation, we first consider various techniques for improving long-tail classification performance which indeed enhance instance segmentation results. We then propose a simple calibration framework to more effectively alleviate classification head bias with a bi-level class balanced sampling approach. Without bells and whistles, it significantly boosts the performance of instance segmentation for tail classes on the recent LVIS dataset and our sampled COCO-LT dataset. Our analysis provides useful insights for solving long-tail instance detection and segmentation problems, and the straightforward \emph{SimCal} method can serve as a simple but strong baseline. With the method we have won the 2019 LVIS challenge. Codes and models are available at https://github.com/twangnh/SimCal.
In recent years, deep network-based methods have continuously refreshed state-of-the-art performance on Salient Object Detection (SOD) task. However, the performance discrepancy caused by different implementation details may conceal the real progress in this task. Making an impartial comparison is required for future researches. To meet this need, we construct a general SALient Object Detection (SALOD) benchmark to conduct a comprehensive comparison among several representative SOD methods. Specifically, we re-implement 14 representative SOD methods by using consistent settings for training. Moreover, two additional protocols are set up in our benchmark to investigate the robustness of existing methods in some limited conditions. In the first protocol, we enlarge the difference between objectness distributions of train and test sets to evaluate the robustness of these SOD methods. In the second protocol, we build multiple train subsets with different scales to validate whether these methods can extract discriminative features from only a few samples. In the above experiments, we find that existing loss functions usually specialized in some metrics but reported inferior results on the others. Therefore, we propose a novel Edge-Aware (EA) loss that promotes deep networks to learn more discriminative features by integrating both pixel- and image-level supervision signals. Experiments prove that our EA loss reports more robust performances compared to existing losses.
Few-shot object detection (FSOD) seeks to detect novel categories with limited data by leveraging prior knowledge from abundant base data. Generalized few-shot object detection (G-FSOD) aims to tackle FSOD without forgetting previously seen base classes and, thus, accounts for a more realistic scenario, where both classes are encountered during test time. While current FSOD methods suffer from catastrophic forgetting, G-FSOD addresses this limitation yet exhibits a performance drop on novel tasks compared to the state-of-the-art FSOD. In this work, we propose a constraint-based finetuning approach (CFA) to alleviate catastrophic forgetting, while achieving competitive results on the novel task without increasing the model capacity. CFA adapts a continual learning method, namely Average Gradient Episodic Memory (A-GEM) to G-FSOD. Specifically, more constraints on the gradient search strategy are imposed from which a new gradient update rule is derived, allowing for better knowledge exchange between base and novel classes. To evaluate our method, we conduct extensive experiments on MS-COCO and PASCAL-VOC datasets. Our method outperforms current FSOD and G-FSOD approaches on the novel task with minor degeneration on the base task. Moreover, CFA is orthogonal to FSOD approaches and operates as a plug-and-play module without increasing the model capacity or inference time.
In the domain of intelligent transportation systems (ITS), collaborative perception has emerged as a promising approach to overcome the limitations of individual perception by enabling multiple agents to exchange information, thus enhancing their situational awareness. Collaborative perception overcomes the limitations of individual sensors, allowing connected agents to perceive environments beyond their line-of-sight and field of view. However, the reliability of collaborative perception heavily depends on the data aggregation strategy and communication bandwidth, which must overcome the challenges posed by limited network resources. To improve the precision of object detection and alleviate limited network resources, we propose an intermediate collaborative perception solution in the form of a graph attention network (GAT). The proposed approach develops an attention-based aggregation strategy to fuse intermediate representations exchanged among multiple connected agents. This approach adaptively highlights important regions in the intermediate feature maps at both the channel and spatial levels, resulting in improved object detection precision. We propose a feature fusion scheme using attention-based architectures and evaluate the results quantitatively in comparison to other state-of-the-art collaborative perception approaches. Our proposed approach is validated using the V2XSim dataset. The results of this work demonstrate the efficacy of the proposed approach for intermediate collaborative perception in improving object detection average precision while reducing network resource usage.
Perception is crucial in the realm of autonomous driving systems, where bird's eye view (BEV)-based architectures have recently reached state-of-the-art performance. The desirability of self-supervised representation learning stems from the expensive and laborious process of annotating 2D and 3D data. Although previous research has investigated pretraining methods for both LiDAR and camera-based 3D object detection, a unified pretraining framework for multimodal BEV perception is missing. In this study, we introduce CALICO, a novel framework that applies contrastive objectives to both LiDAR and camera backbones. Specifically, CALICO incorporates two stages: point-region contrast (PRC) and region-aware distillation (RAD). PRC better balances the region- and scene-level representation learning on the LiDAR modality and offers significant performance improvement compared to existing methods. RAD effectively achieves contrastive distillation on our self-trained teacher model. CALICO's efficacy is substantiated by extensive evaluations on 3D object detection and BEV map segmentation tasks, where it delivers significant performance improvements. Notably, CALICO outperforms the baseline method by 10.5% and 8.6% on NDS and mAP. Moreover, CALICO boosts the robustness of multimodal 3D object detection against adversarial attacks and corruption. Additionally, our framework can be tailored to different backbones and heads, positioning it as a promising approach for multimodal BEV perception.
In this paper, we concern with the problem of how to automatically extract the steps that compose real-life hand activities. This is a key competence towards processing, monitoring and providing video guidance in Mixed Reality systems. We use egocentric vision to observe hand-object interactions in real-world tasks and automatically decompose a video into its constituent steps. Our approach combines hand-object interaction (HOI) detection, object similarity measurement and a finite state machine (FSM) representation to automatically edit videos into steps. We use a combination of Convolutional Neural Networks (CNNs) and the FSM to discover, edit cuts and merge segments while observing real hand activities. We evaluate quantitatively and qualitatively our algorithm on two datasets: the GTEA\cite{li2015delving}, and a new dataset we introduce for Chinese Tea making. Results show our method is able to segment hand-object interaction videos into key step segments with high levels of precision.
멀티모달 대규모 언어 모델(MLLM)은 주로 시각적 질의응답, 캡셔닝, 요약과 같은 자유 형식의 비전-언어(vision-language) 태스크를 중심으로 평가되어 왔습니다. 그러나 실제 활용 영역은 더 구조화된 컴퓨터 비전(CV) 환경으로 빠르게 확장되고 있으며, 사용자는 대규모 에이전트(agent) 또는 의사결정 시스템 내에서 객체 검출(object detection)과 같이 위치 특정(localization) 중심의 작업을 수행하도록 모델(model)에 프롬프트(prompt)를 제공하는 경우가 많습니다. 이러한 변화에도 불구하고, 현재 이러한 역량을 대규모로 체계적으로 평가할 수 있는 표준화된 벤치마크는 부재한 상황입니다. 본 연구에서는 범용 MLLM의 프롬프트 기반 위치 특정 능력을 평가하기 위해 특별히 설계된 최초의 종합적인 벤치마크를 소개합니다. 제안하는 벤치마크는 객체 검출, 지칭 표현 검출(referring expression detection), 인스턴스 수준 검출, 비디오 기반 검출의 네 가지 핵심 작업 범주를 아우릅니다. 일관되고 공정한 평가를 위해 입력을 표준화하고, 파싱 가능한 경계 상자(bounding box) 출력을 강제하며, 태스크 전반에 걸쳐 투명한 평가 프로토콜을 정의하는 통합 프레임워크를 개발했습니다. 이 평가 제품군을 활용하여 다양한 오픈소스 및 상용 MLLM을 평가하고, 성능과 한계에 대한 심층적인 분석을 제공합니다. 정확도 외에도 출력 포맷 사양을 준수하는 모델의 능력을 조사하여, 현재 시스템들이 포맷 제약에 매우 민감하며 사소한 변형에도 일반화하지 못하는 경우가 많음을 보여줍니다. 본 연구 결과는 위치 특정 환경에서 최첨단 MLLM의 강점과 약점을 모두 조명하며, 멀티모달 모델 설계 및 평가를 개선하기 위한 중요한 방향성을 제시합니다.
Aiming at facilitating a real-world, ever-evolving and scalable autonomous driving system, we present a large-scale dataset for standardizing the evaluation of different self-supervised and semi-supervised approaches by learning from raw data, which is the first and largest dataset to date. Existing autonomous driving systems heavily rely on `perfect' visual perception models (i.e., detection) trained using extensive annotated data to ensure safety. However, it is unrealistic to elaborately label instances of all scenarios and circumstances (i.e., night, extreme weather, cities) when deploying a robust autonomous driving system. Motivated by recent advances of self-supervised and semi-supervised learning, a promising direction is to learn a robust detection model by collaboratively exploiting large-scale unlabeled data and few labeled data. Existing datasets either provide only a small amount of data or covers limited domains with full annotation, hindering the exploration of large-scale pre-trained models. Here, we release a Large-Scale 2D Self/semi-supervised Object Detection dataset for Autonomous driving, named as SODA10M, containing 10 million unlabeled images and 20K images labeled with 6 representative object categories. To improve diversity, the images are collected within 27833 driving hours under different weather conditions, periods and location scenes of 32 different cities. We provide extensive experiments and deep analyses of existing popular self/semi-supervised approaches, and give some interesting findings in autonomous driving scope. Experiments show that SODA10M can serve as a promising pre-training dataset for different self-supervised learning methods, which gives superior performance when fine-tuning with different downstream tasks (i.e., detection, semantic/instance segmentation) in autonomous driving domain. More information can refer to https://soda-2d.github.io.
Depth Estimation and Object Detection Recognition play an important role in autonomous driving technology under the guidance of deep learning artificial intelligence. We propose a hybrid structure called RealNet: a co-design method combining the model-streamlined recognition algorithm, the depth estimation algorithm with information fusion, and deploying them on the Jetson-Nano for unmanned vehicles with monocular vision sensors. We use ROS for experiment. The method proposed in this paper is suitable for mobile platforms with high real-time request. Innovation of our method is using information fusion to compensate the problem of insufficient frame rate of output image, and improve the robustness of target detection and depth estimation under monocular vision.Object Detection is based on YOLO-v5. We have simplified the network structure of its DarkNet53 and realized a prediction speed up to 0.01s. Depth Estimation is based on the VNL Depth Estimation, which considers multiple geometric constraints in 3D global space. It calculates the loss function by calculating the deviation of the virtual normal vector VN and the label, which can obtain deeper depth information. We use PnP fusion algorithm to solve the problem of insufficient frame rate of depth map output. It solves the motion estimation depth from three-dimensional target to two-dimensional point based on corner feature matching, which is faster than VNL calculation. We interpolate VNL output and PnP output to achieve information fusion. Experiments show that this can effectively eliminate the jitter of depth information and improve robustness. At the control end, this method combines the results of target detection and depth estimation to calculate the target position, and uses a pure tracking control algorithm to track it.
For the safety of the traveling public, the Transportation Security Administration (TSA) operates security checkpoints at airports in the United States, seeking to keep dangerous items off airplanes. At these checkpoints, the TSA employs a fleet of X-ray scanners, such as the Rapiscan 620DV, so Transportation Security Officers (TSOs) can inspect the contents of carry-on possessions. However, identifying and locating all potential threats can be a challenging task. As a result, the TSA has taken a recent interest in deep learning-based automated detection algorithms that can assist TSOs. In a collaboration funded by the TSA, we collected a sizable new dataset of X-ray scans with a diverse set of threats in a wide array of contexts, trained several deep convolutional object detection models, and integrated such models into the Rapiscan 620DV, resulting in functional prototypes capable of operating in real time. We show performance of our models on held-out evaluation sets, analyze several design parameters, and demonstrate the potential of such systems for automated detection of threats that can be found in airports.
Open-world (OW) recognition and detection models show strong zero- and few-shot adaptation abilities, inspiring their use as initializations in continual learning methods to improve performance. Despite promising results on seen classes, such OW abilities on unseen classes are largely degenerated due to catastrophic forgetting. To tackle this challenge, we propose an open-world continual object detection task, requiring detectors to generalize to old, new, and unseen categories in continual learning scenarios. Based on this task, we present a challenging yet practical OW-COD benchmark to assess detection abilities. The goal is to motivate OW detectors to simultaneously preserve learned classes, adapt to new classes, and maintain open-world capabilities under few-shot adaptations. To mitigate forgetting in unseen categories, we propose MR-GDINO, a strong, efficient and scalable baseline via memory and retrieval mechanisms within a highly scalable memory pool. Experimental results show that existing continual detectors suffer from severe forgetting for both seen and unseen categories. In contrast, MR-GDINO largely mitigates forgetting with only 0.1% activated extra parameters, achieving state-of-the-art performance for old, new, and unseen categories.
Most existing object detectors suffer from class imbalance problems that hinder balanced performance. In particular, anchor free object detectors have to solve the background imbalance problem due to detection in a per-pixel prediction fashion as well as foreground imbalance problem simultaneously. In this work, we propose Balance-oriented focal loss that can induce balanced learning by considering both background and foreground balance comprehensively. This work aims to address imbalance problem in the situation of using a general unbalanced data of non-extreme distribution not including few shot and the focal loss for anchor free object detector. We use a batch-wise alpha-balanced variant of the focal loss to deal with this imbalance problem elaborately. It is a simple and practical solution using only re-weighting for general unbalanced data. It does require neither additional learning cost nor structural change during inference and grouping classes is also unnecessary. Through extensive experiments, we show the performance improvement for each component and analyze the effect of linear scheduling when using re-weighting for the loss. By improving the focal loss in terms of balancing foreground classes, our method achieves AP gains of +1.2 in MS-COCO for the anchor free real-time detector.
Cooperative perception via communication among intelligent traffic agents has great potential to improve the safety of autonomous driving. However, limited communication bandwidth, localization errors and asynchronized capturing time of sensor data, all introduce difficulties to the data fusion of different agents. To some extend, previous works have attempted to reduce the shared data size, mitigate the spatial feature misalignment caused by localization errors and communication delay. However, none of them have considered the asynchronized sensor ticking times, which can lead to dynamic object misplacement of more than one meter during data fusion. In this work, we propose Time-Aligned COoperative Object Detection (TA-COOD), for which we adapt widely used dataset OPV2V and DairV2X with considering asynchronous LiDAR sensor ticking times and build an efficient fully sparse framework with modeling the temporal information of individual objects with query-based techniques. The experiment results confirmed the superior efficiency of our fully sparse framework compared to the state-of-the-art dense models. More importantly, they show that the point-wise observation timestamps of the dynamic objects are crucial for accurate modeling the object temporal context and the predictability of their time-related locations. The official code is available at \url{https://github.com/YuanYunshuang/CoSense3D}.
Localization Quality Estimation (LQE) is crucial and popular in the recent advancement of dense object detectors since it can provide accurate ranking scores that benefit the Non-Maximum Suppression processing and improve detection performance. As a common practice, most existing methods predict LQE scores through vanilla convolutional features shared with object classification or bounding box regression. In this paper, we explore a completely novel and different perspective to perform LQE -- based on the learned distributions of the four parameters of the bounding box. The bounding box distributions are inspired and introduced as "General Distribution" in GFLV1, which describes the uncertainty of the predicted bounding boxes well. Such a property makes the distribution statistics of a bounding box highly correlated to its real localization quality. Specifically, a bounding box distribution with a sharp peak usually corresponds to high localization quality, and vice versa. By leveraging the close correlation between distribution statistics and the real localization quality, we develop a considerably lightweight Distribution-Guided Quality Predictor (DGQP) for reliable LQE based on GFLV1, thus producing GFLV2. To our best knowledge, it is the first attempt in object detection to use a highly relevant, statistical representation to facilitate LQE. Extensive experiments demonstrate the effectiveness of our method. Notably, GFLV2 (ResNet-101) achieves 46.2 AP at 14.6 FPS, surpassing the previous state-of-the-art ATSS baseline (43.6 AP at 14.6 FPS) by absolute 2.6 AP on COCO {\tt test-dev}, without sacrificing the efficiency both in training and inference. Code will be available at https://github.com/implus/GFocalV2.
The task of LiDAR-based 3D Open-Vocabulary Detection (3D OVD) requires the detector to learn to detect novel objects from point clouds without off-the-shelf training labels. Previous methods focus on the learning of object-level representations and ignore the scene-level information, thus it is hard to distinguish objects with similar classes. In this work, we propose a Global-Local Collaborative Reason and Debate with PSL (GLRD) framework for the 3D OVD task, considering both local object-level information and global scene-level information. Specifically, LLM is utilized to perform common sense reasoning based on object-level and scene-level information, where the detection result is refined accordingly. To further boost the LLM's ability of precise decisions, we also design a probabilistic soft logic solver (OV-PSL) to search for the optimal solution, and a debate scheme to confirm the class of confusable objects. In addition, to alleviate the uneven distribution of classes, a static balance scheme (SBC) and a dynamic balance scheme (DBC) are designed. In addition, to reduce the influence of noise in data and training, we further propose Reflected Pseudo Labels Generation (RPLG) and Background-Aware Object Localization (BAOL). Extensive experiments conducted on ScanNet and SUN RGB-D demonstrate the superiority of GLRD, where absolute improvements in mean average precision are $+2.82\%$ on SUN RGB-D and $+3.72\%$ on ScanNet in the partial open-vocabulary setting. In the full open-vocabulary setting, the absolute improvements in mean average precision are $+4.03\%$ on ScanNet and $+14.11\%$ on SUN RGB-D.
Monocular multi-object detection and localization in 3D space has been proven to be a challenging task. The MoNet3D algorithm is a novel and effective framework that can predict the 3D position of each object in a monocular image and draw a 3D bounding box for each object. The MoNet3D method incorporates prior knowledge of the spatial geometric correlation of neighbouring objects into the deep neural network training process to improve the accuracy of 3D object localization. Experiments on the KITTI dataset show that the accuracy for predicting the depth and horizontal coordinates of objects in 3D space can reach 96.25\% and 94.74\%, respectively. Moreover, the method can realize the real-time image processing at 27.85 FPS, showing promising potential for embedded advanced driving-assistance system applications. Our code is publicly available at https://github.com/CQUlearningsystemgroup/YicongPeng.
We present the training recipe and results of scaling up PaLI-X, a multilingual vision and language model, both in terms of size of the components and the breadth of its training task mixture. Our model achieves new levels of performance on a wide-range of varied and complex tasks, including multiple image-based captioning and question-answering tasks, image-based document understanding and few-shot (in-context) learning, as well as object detection, video question answering, and video captioning. PaLI-X advances the state-of-the-art on most vision-and-language benchmarks considered (25+ of them). Finally, we observe emerging capabilities, such as complex counting and multilingual object detection, tasks that are not explicitly in the training mix.
Object detectors frequently encounter significant performance degradation when confronted with domain gaps between collected data (source domain) and data from real-world applications (target domain). To address this task, numerous unsupervised domain adaptive detectors have been proposed, leveraging carefully designed feature alignment techniques. However, these techniques primarily align instance-level features in a class-agnostic manner, overlooking the differences between extracted features from different categories, which results in only limited improvement. Furthermore, the scope of current alignment modules is often restricted to a limited batch of images, failing to learn the entire dataset-level cues, thereby severely constraining the detector's generalization ability to the target domain. To this end, we introduce a strong DETR-based detector named Domain Adaptive detection TRansformer (DATR) for unsupervised domain adaptation of object detection. Firstly, we propose the Class-wise Prototypes Alignment (CPA) module, which effectively aligns cross-domain features in a class-aware manner by bridging the gap between object detection task and domain adaptation task. Then, the designed Dataset-level Alignment Scheme (DAS) explicitly guides the detector to achieve global representation and enhance inter-class distinguishability of instance-level features across the entire dataset, which spans both domains, by leveraging contrastive learning. Moreover, DATR incorporates a mean-teacher based self-training framework, utilizing pseudo-labels generated by the teacher model to further mitigate domain bias. Extensive experimental results demonstrate superior performance and generalization capabilities of our proposed DATR in multiple domain adaptation scenarios. Code is released at https://github.com/h751410234/DATR.
CNN과 SSM의 주파수 선호 특성이 본질적으로 상호 보완적임을 스펙트럼 분석으로 규명하고, Liquid Neural Networks의 연속 시간 동역학에서 영감을 받아 VMamba와 ConvNeXt 두 백본의 이질적 표현을 동적으로 융합하는 LFNet을 제안한다. RGB, RGB-D, RGB-T, VSOD, VDT 등 다섯 가지 일반 현저 객체 검출(SOD) 태스크에서 최고 성능을 달성하며 검출 정확도와 모델 효율성 간의 우수한 트레이드오프를 보인다.
Real-world object detection models should be cheap and accurate. Knowledge distillation (KD) can boost the accuracy of a small, cheap 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.