The presence of Non-Line-of-Sight (NLoS) blind spots resulting from roadside parking in urban environments poses a significant challenge to road safety, particularly due to the sudden emergence of pedestrians. mmWave technology leverages diffraction and reflection to observe NLoS regions, and recent studies have demonstrated its potential for detecting obscured objects. However, existing approaches predominantly rely on predefined spatial information or assume simple wall reflections, thereby limiting their generalizability and practical applicability. A particular challenge arises in scenarios where pedestrians suddenly appear from between parked vehicles, as these parked vehicles act as temporary spatial obstructions. Furthermore, since parked vehicles are dynamic and may relocate over time, spatial information obtained from satellite maps or other predefined sources may not accurately reflect real-time road conditions, leading to erroneous sensor interpretations. To address this limitation, we propose an NLoS pedestrian localization framework that integrates monocular camera image with 2D radar point cloud (PCD) data. The proposed method initially detects parked vehicles through image segmentation, estimates depth to infer approximate spatial characteristics, and subsequently refines this information using 2D radar PCD to achieve precise spatial inference. Experimental evaluations conducted in real-world urban road environments demonstrate that the proposed approach enhances early pedestrian detection and contributes to improved road safety. Supplementary materials are available at https://hiyeun.github.io/NLoS/.

Pedestrians Localization in Non-Line-of-Sight (NLoS) regions within urban environments poses a significant challenge for autonomous driving systems. While mmWave radar has demonstrated potential for detecting objects in such scenarios, the 2D radar point cloud (PCD) data is susceptible to distortions caused by multipath reflections, making accurate spatial inference difficult. Additionally, although camera images provide high-resolution visual information, they lack depth perception and cannot directly observe objects in NLoS regions. In this paper, we propose a novel framework that interprets radar PCD through road layout inferred from camera for localization of NLoS pedestrians. The proposed method leverages visual information from the camera to interpret 2D radar PCD, enabling spatial scene reconstruction. The effectiveness of the proposed approach is validated through experiments conducted using a radar-camera system mounted on a real vehicle. The localization performance is evaluated using a dataset collected in outdoor NLoS driving environments, demonstrating the practical applicability of the method.

Autonomous vehicles are increasingly utilized in diverse industries, relying heavily on perception systems to interpret their surroundings for decision-making and control. While Line-of-Sight perception technologies have advanced significantly, Non-Line-of-Sight (NLoS) perception remains a critical challenge. Current systems struggle to detect objects in NLoS scenarios, such as pedestrians or vehicles suddenly appearing from behind obstacles, leading to accidents, particularly at narrow T-junctions in urban environments. To address this, mmWave radar has emerged as a promising sensor for NLoS perception due to its ability to capture reflections and estimate the location of dynamic objects in occluded areas. However, previous researches are limited to controlled settings or single objects, with challenges like multipath reflections requiring precise spatial analysis for real-world use. In this paper, we propose a localization method for multi-dynamic NLoS pedestrians using ray tracing on 2D radar point clouds obtained from mm Wave radar in outdoor environments. The approach involves inferring spatial information from static points, performing ray tracing for dynamic points, and applying noise filtering and clustering to estimate pedestrian locations. Validation on a custom-built test bed demonstrates the effectiveness of the method, establishing a foundation for advanced NLoS perception technologies in real-world driving.

In this paper, we propose SPACE, a novel anomaly detection methodology that integrates a Feature Encoder (FE) into the structure of the Student-Teacher method. The proposed method has two key elements: Spatial Consistency regularization Loss (SCL) and Feature converter Module (FM). SCL prevents overfitting in student models by avoiding excessive imitation of the teacher model. Simultaneously, it facilitates the expansion of normal data features by steering clear of abnormal areas generated through data augmentation. This dual functionality ensures a robust boundary between normal and abnormal data. The FM prevents the learning of ambiguous information from the FE. This protects the learned features and enables more effective detection of structural and logical anomalies. Through these elements, SPACE is available to minimize the influence of the FE while integrating various data augmentations. In this study, we evaluated the proposed method on the MVTec LOCO, MVTec AD, and VisA datasets. Experimental results, through qualitative evaluation, demonstrate the superiority of detection and efficiency of each module compared to state-of-the-art methods.

Abstract Annotating the dataset with high-quality labels is crucial for deep networks’ performance, but in real-world scenarios, the labels are often contaminated by noise. To address this, some methods were recently proposed to automatically split clean and noisy labels among training data, and learn a semi-supervised learner in a Learning with Noisy Labels (LNL) framework. However, they leverage a handcrafted module for clean-noisy label splitting, which induces a confirmation bias in the semi-supervised learning phase and limits the performance. In this paper, for the first time, we present a learnable module for clean-noisy label splitting, dubbed SplitNet, and a novel LNL framework which complementarily trains the SplitNet and main network for the LNL task. We also propose to use a dynamic threshold based on split confidence by SplitNet to optimize the semi-supervised learner better. To enhance SplitNet training, we further present a risk hedging method. Our proposed method performs at a state-of-the-art level, especially in high noise ratio settings on various LNL benchmarks.

Extracting and exploiting edge information is important for various computer vision tasks. According to the physical characteristics, edges are further categorized into reflectance, illumination, normal, and depth discontinuities. Previous studies for edge discontinuity classification have achieved impressive performance in extracting discontinuities, but the capacity of classification remains limited. This paper proposes TriDecTr, a novel network comprising a transformer-based encoder to improve semantic understanding and a tri-directional decoder to explore relationships among categories. Extensive experiments demonstrate that TriDecTr achieves state-of-the-art performance on the BSDS-RIND dataset with 0.531 in ODS, 0.571 in OIS, and 0.461 in AP. Moreover, TriDecTr significantly narrows the performance gap between illumination edges and the other categories from 0.191 to 0.118 in ODS.

Recent advances in deep learning have significantly improved the performance of various computer vision applications. However, discovering novel categories in an incremental learning scenario remains a challenging problem due to the lack of prior knowledge about the number and nature of new categories. Existing methods for novel category discovery are limited by their reliance on labeled datasets and prior knowledge about the number of novel categories and the proportion of novel samples in the batch. To address the limitations and more accurately reflect real-world scenarios, in this paper, we propose a novel unsupervised class incremental learning approach for discovering novel categories on unlabeled sets without prior knowledge. The proposed method fine-tunes the feature extractor and proxy anchors on labeled sets, then splits samples into old and novel categories and clusters on the unlabeled dataset. Furthermore, the proxy anchors-based exemplar generates representative category vectors to mitigate catastrophic forgetting. Experimental results demonstrate that our proposed approach outperforms the state-of-the-art methods on fine-grained datasets under real-world scenarios.

Establishing dense correspondences across semantically similar images remains a challenging task due to the significant intra-class variations and background clutters. Traditionally, a supervised learning was used for training the models, which required tremendous manually-labeled data, while some methods suggested a self-supervised or weakly-supervised learning to mitigate the reliance on the labeled data, but with limited performance. In this paper, we present a simple, but effective solution for semantic correspondence that learns the networks in a semi-supervised manner by supplementing few ground-truth correspondences via utilization of a large amount of confident correspondences as pseudo-labels, called SemiMatch. Specifically, our framework generates the pseudo-labels using the model's prediction itself between source and weakly-augmented target, and uses pseudo-labels to learn the model again between source and strongly-augmented target, which improves the robustness of the model. We also present a novel confidence measure for pseudo-labels and data augmentation tailored for semantic correspondence. In experiments, SemiMatch achieves state-of-the-art performance on various benchmarks.

3D point cloud data is used for 3D model acquisition, geometry processing and 3D inspection. Registration of 3D point cloud data is crucial for each field. The difference between 2D image registration and 3D point cloud registration is that the latter requires several things to be considered: translation on each plane, rotation, tilt and etc. This paper describes a method of registering 3D point cloud data with noise. The relationship between the two sets of 3D point cloud data can be obtained by Affine transformation. In order to calculate 3D Affine transformation matrix, corresponding points are required. To find the corresponding points, we use the height map which is projected from 3D point cloud data onto XY plane. We formulate the height map matching as a cost function and estimate the corresponding points. To find the proper 3D Affine transformation matrix, we formulate a cost function which uses the relationship of the corresponding points. Also the proper 3D Affine transformation matrix can be calculated by minimizing the cost function. The experimental results show that the proposed method can be applied to various objects and gives better performance than the previous work.