While object detection methods traditionally make use of pixel-level masks or bounding boxes, alternative representations such as polygons or active contours have recently emerged. Among them, methods based on the regression of Fourier or Chebyshev coefficients have shown high potential on freeform objects. By defining object shapes as polar functions, they are however limited to star-shaped domains. We address this issue with SCR: a method that captures resolution-free object contours as complex periodic functions. The method offers a good compromise between accuracy and compactness thanks to the design of efficient geometric shape priors. We benchmark SCR on the popular COCO 2017 instance segmentation dataset, and show its competitiveness against existing algorithms in the field. In addition, we design a compact version of our network, which we benchmark on embedded hardware with a wide range of power targets, achieving up to real-time performance.
The promise of self-supervised learning (SSL) is to leverage large amounts of unlabeled data to solve complex tasks. While there has been excellent progress with simple, image-level learning, recent methods have shown the advantage of including knowledge of image structure. However, by introducing hand-crafted image segmentations to define regions of interest, or specialized augmentation strategies, these methods sacrifice the simplicity and generality that makes SSL so powerful. Instead, we propose a self-supervised learning paradigm that discovers this image structure by itself. Our method, Odin, couples object discovery and representation networks to discover meaningful image segmentations without any supervision. The resulting learning paradigm is simpler, less brittle, and more general, and achieves state-of-the-art transfer learning results for object detection and instance segmentation on COCO, and semantic segmentation on PASCAL and Cityscapes, while strongly surpassing supervised pre-training for video segmentation on DAVIS.
Existing Visual Question Answering (VQA) models have explored various visual relationships between objects in the image to answer complex questions, which inevitably introduces irrelevant information brought by inaccurate object detection and text grounding. To address the problem, we propose a Question-Driven Graph Fusion Network (QD-GFN). It first models semantic, spatial, and implicit visual relations in images by three graph attention networks, then question information is utilized to guide the aggregation process of the three graphs, further, our QD-GFN adopts an object filtering mechanism to remove question-irrelevant objects contained in the image. Experiment results demonstrate that our QD-GFN outperforms the prior state-of-the-art on both VQA 2.0 and VQA-CP v2 datasets. Further analysis shows that both the novel graph aggregation method and object filtering mechanism play a significant role in improving the performance of the model.
Mobile autonomy relies on the precise perception of dynamic environments. Robustly tracking moving objects in 3D world thus plays a pivotal role for applications like trajectory prediction, obstacle avoidance, and path planning. While most current methods utilize LiDARs or cameras for Multiple Object Tracking (MOT), the capabilities of 4D imaging radars remain largely unexplored. Recognizing the challenges posed by radar noise and point sparsity in 4D radar data, we introduce RaTrack, an innovative solution tailored for radar-based tracking. Bypassing the typical reliance on specific object types and 3D bounding boxes, our method focuses on motion segmentation and clustering, enriched by a motion estimation module. Evaluated on the View-of-Delft dataset, RaTrack showcases superior tracking precision of moving objects, largely surpassing the performance of the state of the art. We release our code and model at https://github.com/LJacksonPan/RaTrack.
Anomaly detection from a driver's perspective when driving is important to autonomous vehicles. As a part of Advanced Driver Assistance Systems (ADAS), it can remind the driver about dangers timely. Compared with traditional studied scenes such as the university campus and market surveillance videos, it is difficult to detect abnormal event from a driver's perspective due to camera waggle, abidingly moving background, drastic change of vehicle velocity, etc. To tackle these specific problems, this paper proposes a spatial localization constrained sparse coding approach for anomaly detection in traffic scenes, which firstly measures the abnormality of motion orientation and magnitude respectively and then fuses these two aspects to obtain a robust detection result. The main contributions are threefold: 1) This work describes the motion orientation and magnitude of the object respectively in a new way, which is demonstrated to be better than the traditional motion descriptors. 2) The spatial localization of object is taken into account of the sparse reconstruction framework, which utilizes the scene's structural information and outperforms the conventional sparse coding methods. 3) Results of motion orientation and magnitude are adaptively weighted and fused by a Bayesian model, which makes the proposed method more robust and handle more kinds of abnormal events. The efficiency and effectiveness of the proposed method are validated by testing on nine difficult video sequences captured by ourselves. Observed from the experimental results, the proposed method is more effective and efficient than the popular competitors, and yields a higher performance.
Event cameras capture sparse brightness changes with high temporal resolution and high dynamic range, compensating for the deficiencies of the conventional RGB frames. However, previous multi-modal fusion techniques typically fail to handle the inherent heterogeneity between RGB frames and event streams, thus easily leading to noise amplification or redundant feature integration during cross-modal fusion. In this paper, we propose a Cross-Modal information inTeraction transFormer, coined as CMTFormer, which hierarchically integrates RGB and event information to achieve efficient and stable multimodal collaboration. Specifically, we design a shallow-to-deep information interaction scheme. In the shallow stage, we present the Shallow Alignment Module (SAM) to achieve an efficient fusion of RGB and event low-level features, which mitigates attribute disparities and prevents noisy information. In the middle stage, we devise the Cross-modal Enhancement Module (CEM) that utilizes texture and edge information to produce mutually reinforced middle-level features. In the deep stage, we present the Learnable Deep Fusion Module (LDFM) which performs high-level information aggregation through learnable weights, thus enabling the network to adaptively fuse RGB and event clues. A Spatial Prior Module is further designed to utilize global spatial information to enhance localization accuracy. Extensive experiments are conducted on two prevalent event-based object detection benchmarks, i.e., DSEC-Detection and PKU-DAVIS-SOD. Our CMTFormer consistently surpasses the detection counterparts in both uni-modal and multi-modal settings, strongly demonstrating the effectiveness of our paradigm. Codes will be available upon publication.
Anomalous sound detection (ASD) is the task of identifying whether the sound emitted from an object is normal or anomalous. In some cases, early detection of this anomaly can prevent several problems. This article presents a Systematic Review (SR) about studies related to Anamolous Sound Detection using Machine Learning (ML) techniques. This SR was conducted through a selection of 31 (accepted studies) studies published in journals and conferences between 2010 and 2020. The state of the art was addressed, collecting data sets, methods for extracting features in audio, ML models, and evaluation methods used for ASD. The results showed that the ToyADMOS, MIMII, and Mivia datasets, the Mel-frequency cepstral coefficients (MFCC) method for extracting features, the Autoencoder (AE) and Convolutional Neural Network (CNN) models of ML, the AUC and F1-score evaluation methods were most cited.
Detection and segmentation of objects in overheard imagery is a challenging task. The variable density, random orientation, small size, and instance-to-instance heterogeneity of objects in overhead imagery calls for approaches distinct from existing models designed for natural scene datasets. Though new overhead imagery datasets are being developed, they almost universally comprise a single view taken from directly overhead ("at nadir"), failing to address a critical variable: look angle. By contrast, views vary in real-world overhead imagery, particularly in dynamic scenarios such as natural disasters where first looks are often over 40 degrees off-nadir. This represents an important challenge to computer vision methods, as changing view angle adds distortions, alters resolution, and changes lighting. At present, the impact of these perturbations for algorithmic detection and segmentation of objects is untested. To address this problem, we present an open source Multi-View Overhead Imagery dataset, termed SpaceNet MVOI, with 27 unique looks from a broad range of viewing angles (-32.5 degrees to 54.0 degrees). Each of these images cover the same 665 square km geographic extent and are annotated with 126,747 building footprint labels, enabling direct assessment of the impact of viewpoint perturbation on model performance. We benchmark multiple leading segmentation and object detection models on: (1) building detection, (2) generalization to unseen viewing angles and resolutions, and (3) sensitivity of building footprint extraction to changes in resolution. We find that state of the art segmentation and object detection models struggle to identify buildings in off-nadir imagery and generalize poorly to unseen views, presenting an important benchmark to explore the broadly relevant challenge of detecting small, heterogeneous target objects in visually dynamic contexts.
Small target detection in UAV imagery faces significant challenges such as scale variations, dense distribution, and the dominance of small targets. Existing algorithms rely on manually designed components, and general-purpose detectors are not optimized for UAV images, making it difficult to balance accuracy and complexity. To address these challenges, this paper proposes an end-to-end object detection framework, UFO-DETR, which integrates an LSKNet-based backbone network to optimize the receptive field and reduce the number of parameters. By combining the DAttention and AIFI modules, the model flexibly models multi-scale spatial relationships, improving multi-scale target detection performance. Additionally, the DynFreq-C3 module is proposed to enhance small target detection capability through cross-space frequency feature enhancement. Experimental results show that, compared to RT-DETR-L, the proposed method offers significant advantages in both detection performance and computational efficiency, providing an efficient solution for UAV edge computing.
Deep SORT\cite{wojke2017simple} is a tracking-by-detetion approach to multiple object tracking with a detector and a RE-ID model. Both separately training and inference with the two model is time-comsuming. In this paper, we unify the detector and RE-ID model into an end-to-end network, by adding an additional track branch for tracking in Faster RCNN architecture. With a unified network, we are able to train the whole model end-to-end with multi loss, which has shown much benefit in other recent works. The RE-ID model in Deep SORT needs to use deep CNNs to extract feature map from detected object images, However, track branch in our proposed network straight make use of the RoI feature vector in Faster RCNN baseline, which reduced the amount of calculation. Since the single image lacks the same object which is necessary when we use the triplet loss to optimizer the track branch, we concatenate the neighbouring frames in a video to construct our training dataset. We have trained and evaluated our model on AIC19 vehicle tracking dataset, experiment shows that our model with resnet101 backbone can achieve 57.79 \% mAP and track vehicle well.
Adversarial examples have gained tons of attention in recent years. Many adversarial attacks have been proposed to attack image classifiers, but few work shift attention to object detectors. In this paper, we propose Sparse Adversarial Attack (SAA) which enables adversaries to perform effective evasion attack on detectors with bounded \emph{l$_{0}$} norm perturbation. We select the fragile position of the image and designed evasion loss function for the task. Experiment results on YOLOv4 and FasterRCNN reveal the effectiveness of our method. In addition, our SAA shows great transferability across different detectors in the black-box attack setting. Codes are available at \emph{https://github.com/THUrssq/Tianchi04}.
Dominated point cloud-based 3D object detectors in autonomous driving scenarios rely heavily on the huge amount of accurately labeled samples, however, 3D annotation in the point cloud is extremely tedious, expensive and time-consuming. To reduce the dependence on large supervision, semi-supervised learning (SSL) based approaches have been proposed. The Pseudo-Labeling methodology is commonly used for SSL frameworks, however, the low-quality predictions from the teacher model have seriously limited its performance. In this work, we propose a new Pseudo-Labeling framework for semi-supervised 3D object detection, by enhancing the teacher model to a proficient one with several necessary designs. First, to improve the recall of pseudo labels, a Spatialtemporal Ensemble (STE) module is proposed to generate sufficient seed boxes. Second, to improve the precision of recalled boxes, a Clusteringbased Box Voting (CBV) module is designed to get aggregated votes from the clustered seed boxes. This also eliminates the necessity of sophisticated thresholds to select pseudo labels. Furthermore, to reduce the negative influence of wrongly pseudo-labeled samples during the training, a soft supervision signal is proposed by considering Box-wise Contrastive Learning (BCL). The effectiveness of our model is verified on both ONCE and Waymo datasets. For example, on ONCE, our approach significantly improves the baseline by 9.51 mAP. Moreover, with half annotations, our model outperforms the oracle model with full annotations on Waymo.
To achieve accurate and low-cost 3D object detection, existing methods propose to benefit camera-based multi-view detectors with spatial cues provided by the LiDAR modality, e.g., dense depth supervision and bird-eye-view (BEV) feature distillation. However, they directly conduct point-to-point mimicking from LiDAR to camera, which neglects the inner-geometry of foreground targets and suffers from the modal gap between 2D-3D features. In this paper, we propose the learning scheme of Target Inner-Geometry from the LiDAR modality into camera-based BEV detectors for both dense depth and BEV features, termed as TiG-BEV. First, we introduce an inner-depth supervision module to learn the low-level relative depth relations between different foreground pixels. This enables the camera-based detector to better understand the object-wise spatial structures. Second, we design an inner-feature BEV distillation module to imitate the high-level semantics of different keypoints within foreground targets. To further alleviate the BEV feature gap between two modalities, we adopt both inter-channel and inter-keypoint distillation for feature-similarity modeling. With our target inner-geometry distillation, TiG-BEV can effectively boost BEVDepth by +2.3% NDS and +2.4% mAP, along with BEVDet by +9.1% NDS and +10.3% mAP on nuScenes val set. Code will be available at https://github.com/ADLab3Ds/TiG-BEV.
High-resolution representations are essential for position-sensitive vision problems, such as human pose estimation, semantic segmentation, and object detection. Existing state-of-the-art frameworks first encode the input image as a low-resolution representation through a subnetwork that is formed by connecting high-to-low resolution convolutions \emph{in series} (e.g., ResNet, VGGNet), and then recover the high-resolution representation from the encoded low-resolution representation. Instead, our proposed network, named as High-Resolution Network (HRNet), maintains high-resolution representations through the whole process. There are two key characteristics: (i) Connect the high-to-low resolution convolution streams \emph{in parallel}; (ii) Repeatedly exchange the information across resolutions. The benefit is that the resulting representation is semantically richer and spatially more precise. We show the superiority of the proposed HRNet in a wide range of applications, including human pose estimation, semantic segmentation, and object detection, suggesting that the HRNet is a stronger backbone for computer vision problems. All the codes are available at~{\url{https://github.com/HRNet}}.
Salient object detection (SOD) has been well studied in recent years, especially using deep neural networks. However, SOD with RGB and RGB-D images is usually treated as two different tasks with different network structures that need to be designed specifically. In this paper, we proposed a unified and efficient structure with a cross-attention context extraction (CRACE) module to address both tasks of SOD efficiently. The proposed CRACE module receives and appropriately fuses two (for RGB SOD) or three (for RGB-D SOD) inputs. The simple unified feature pyramid network (FPN)-like structure with CRACE modules conveys and refines the results under the multi-level supervisions of saliency and boundaries. The proposed structure is simple yet effective; the rich context information of RGB and depth can be appropriately extracted and fused by the proposed structure efficiently. Experimental results show that our method outperforms other state-of-the-art methods in both RGB and RGB-D SOD tasks on various datasets and in terms of most metrics.
One-shot object detection aims at detecting novel objects according to merely one given instance. With extreme data scarcity, current approaches explore various feature fusions to obtain directly transferable meta-knowledge. Yet, their performances are often unsatisfactory. In this paper, we attribute this to inappropriate correlation methods that misalign query-support semantics by overlooking spatial structures and scale variances. Upon analysis, we leverage the attention mechanism and propose a simple but effective architecture named Semantic-aligned Fusion Transformer (SaFT) to resolve these issues. Specifically, we equip SaFT with a vertical fusion module (VFM) for cross-scale semantic enhancement and a horizontal fusion module (HFM) for cross-sample feature fusion. Together, they broaden the vision for each feature point from the support to a whole augmented feature pyramid from the query, facilitating semantic-aligned associations. Extensive experiments on multiple benchmarks demonstrate the superiority of our framework. Without fine-tuning on novel classes, it brings significant performance gains to one-stage baselines, lifting state-of-the-art results to a higher level.
LiDAR-produced point clouds are the major source for most state-of-the-art 3D object detectors. Yet, small, distant, and incomplete objects with sparse or few points are often hard to detect. We present Sparse2Dense, a new framework to efficiently boost 3D detection performance by learning to densify point clouds in latent space. Specifically, we first train a dense point 3D detector (DDet) with a dense point cloud as input and design a sparse point 3D detector (SDet) with a regular point cloud as input. Importantly, we formulate the lightweight plug-in S2D module and the point cloud reconstruction module in SDet to densify 3D features and train SDet to produce 3D features, following the dense 3D features in DDet. So, in inference, SDet can simulate dense 3D features from regular (sparse) point cloud inputs without requiring dense inputs. We evaluate our method on the large-scale Waymo Open Dataset and the Waymo Domain Adaptation Dataset, showing its high performance and efficiency over the state of the arts.
The dominant multi-camera 3D detection paradigm is based on explicit 3D feature construction, which requires complicated indexing of local image-view features via 3D-to-2D projection. Other methods implicitly introduce geometric positional encoding and perform global attention (e.g., PETR) to build the relationship between image tokens and 3D objects. The 3D-to-2D perspective inconsistency and global attention lead to a weak correlation between foreground tokens and queries, resulting in slow convergence. We propose Focal-PETR with instance-guided supervision and spatial alignment module to adaptively focus object queries on discriminative foreground regions. Focal-PETR additionally introduces a down-sampling strategy to reduce the consumption of global attention. Due to the highly parallelized implementation and down-sampling strategy, our model, without depth supervision, achieves leading performance on the large-scale nuScenes benchmark and a superior speed of 30 FPS on a single RTX3090 GPU. Extensive experiments show that our method outperforms PETR while consuming 3x fewer training hours. The code will be made publicly available.
As an increasing amount of image and video content will be analyzed by machines, there is demand for a new codec paradigm that is capable of compressing visual input primarily for the purpose of computer vision inference, while secondarily supporting input reconstruction. In this work, we propose a learned compression architecture that can be used to build such a codec. We introduce a novel variational formulation that explicitly takes feature data relevant to the desired inference task as input at the encoder side. As such, our learned scalable image codec encodes and transmits two disentangled latent representations for object detection and input reconstruction. We note that compared to relevant benchmarks, our proposed scheme yields a more compact latent representation that is specialized for the inference task. Our experiments show that our proposed system achieves a bit rate savings of 40.6% on the primary object detection task compared to the current state-of-the-art, albeit with some degradation in performance for the secondary input reconstruction task.
In this paper, we explore incremental few-shot object detection (iFSD), which incrementally learns novel classes using only a few examples without revisiting base classes. Previous iFSD works achieved the desired results by applying meta-learning. However, meta-learning approaches show insufficient performance that is difficult to apply to practical problems. In this light, we propose a simple fine-tuning-based approach, the Incremental Two-stage Fine-tuning Approach (iTFA) for iFSD, which contains three steps: 1) base training using abundant base classes with the class-agnostic box regressor, 2) separation of the RoI feature extractor and classifier into the base and novel class branches for preserving base knowledge, and 3) fine-tuning the novel branch using only a few novel class examples. We evaluate our iTFA on the real-world datasets PASCAL VOC, COCO, and LVIS. iTFA achieves competitive performance in COCO and shows a 30% higher AP accuracy than meta-learning methods in the LVIS dataset. Experimental results show the effectiveness and applicability of our proposed method.