Hao Hu

In recent years, visual 3D Semantic Scene Completion (SSC) has emerged as a critical perception task for autonomous driving due to its ability to infer complete 3D scene layouts and semantics from single 2D images. However, in real-world traffic scenarios, a significant portion of the scene remains occluded or outside the camera's field of view - a fundamental challenge that existing monocular SSC methods fail to address adequately. To overcome these limitations, we propose Creating the Future SSC (CF-SSC), a novel temporal SSC framework that leverages pseudo-future frame prediction to expand the model's effective perceptual range. Our approach combines poses and depths to establish accurate 3D correspondences, enabling geometrically-consistent fusion of past, present, and predicted future frames in 3D space. Unlike conventional methods that rely on simple feature stacking, our 3D-aware architecture achieves more robust scene completion by explicitly modeling spatial-temporal relationships. Comprehensive experiments on SemanticKITTI [1] and SSCBench-KITTI-360 [2], [3] benchmarks demonstrate state-of-the-art performance, validating the effectiveness of our approach, highlighting our method's ability to improve occlusion reasoning and 3D scene completion accuracy.

Human drivers excel at dynamically prioritizing safety-critical cues through rapid saccadic gaze shifts-an essential capability often lacking in autonomous systems. To address this gap, we propose a human-inspired framework that simulates saccadic gaze behavior to guide visual attention in driving scenarios. Specifically, our Recurrent Perception Network iteratively predicts human-like gaze regions via multiround training, achieving 85% coverage of annotated human gaze points on validation datasets. These gaze maps are then integrated into a novel Perception-Guided Driving Network, which employs cascaded attention refinement to amplify safetycritical features. Experimental results demonstrate substantial improvements over baseline methods, including 19% higher driving scores, 81.1% better traffic signal compliance, 83.3% fewer pedestrian collisions and 88.8% reduction in red light violations. These findings validate that incorporating human perceptual priors enables autonomous systems to adaptively focus on task-relevant regions in a biologically inspired manner. Our approach bridges machine perception with human cognitive strategies, paving the way for safer, human-aligned autonomous driving without sacrificing generalization.

Introduction: Animal-involved scenarios pose significant challenges for autonomous driving systems due to their rarity, unpredictability, and safety-critical nature. Despite their importance, existing vision-language datasets for autonomous driving largely overlook these long-tail situations. Methods: To address this gap, we introduce AniDriveQA, a novel visual question answering (VQA) dataset specifically designed to evaluate vision-language models (VLMs) in driving scenarios involving animals. The dataset is constructed through a scalable pipeline that collects diverse animal-related traffic scenes from internet videos, filters and annotates them using object detection and scene classification models, and generates multi-task VQA labels with a large vision-language model. AniDriveQA includes three key task types: scene description, animal description, and driving suggestion. Results: For evaluation, a hybrid scheme was employed that combined classification accuracy for structured tasks with LLM-based scoring for open-ended responses. Extensive experiments on various open-source VLMs revealed large performance disparities across models and task types. Discussion: The experimental results demonstrate that AniDriveQA effectively exposes the limitations of current VLMs in rare yet safety-critical autonomous driving scenarios. The dataset provides a valuable diagnostic benchmark for advancing reasoning, perception, and decision-making capabilities in future vision-language models.

High-quality pseudo-labels are critical for guiding learning in semi-supervised change detection (SSCD). Recently, many SSCD methods based on consistency regularization (CR) have achieved advanced performance. These methods typically generate pseudo-labels by setting a fixed threshold. However, this strategy struggles to generate pseudo-labels with high-quality boundary details. To this end, we propose a novel SSCD method boundary refinement teacher (BRT) to enhance the boundary quality of pseudo-labels. A bi-temporal image boundary refinement (BIBR) module is designed to uncover boundary details in unlabeled images at first. BIBR explores the boundary characteristics of pseudo-label by extracting the boundary blocks from its change map and re-delineates the boundary decision points with their magnified views. Then, a stable teacher parameter update (STPU) module is devised to sustain the semi-supervised learning state steady, avoiding frequent updates to the teacher model parameters. These stable updates to teacher model parameters provide continuous and high-quality guidance, preventing pseudo-label fluctuations from disrupting the student model’s acquisition of new knowledge. Extensive experiments are conducted on three commonly used change detection datasets, encompassing buildings and multiple categories, covering SSCD settings, boundary metrics, and detailed ablation studies. Our results show that simply enhancing the boundary quality of the pseudo-labels allows BRT to consistently deliver state-of-the-art (SOTA) performance in SSCD. The code is available at: https://github.com/yoghurts-sy/BRT.

Deep convolutional neural networks (CNNs) have achieved remarkable success in various computer vision tasks. However, the lack of interpretability in these models has raised concerns and hindered their widespread adoption in critical domains. Generating activation maps that highlight the regions contributing to the CNN's decision has emerged as a popular approach to visualize and interpret these models. Nevertheless, existing methods often produce activation maps contaminated with irrelevant background noise or incomplete object activation, limiting their effectiveness in providing meaningful explanations. To address this challenge, we propose Union Class Activation Mapping (UnionCAM), an innovative visual interpretation framework that generates high-quality class activation maps (CAMs) through a novel three-step approach. UnionCAM introduces a weighted fusion strategy that adaptively combines multiple CAMs to create more informative and comprehensive activation maps. First, the denoising module removes background noise from CAMs by using adaptive thresholding. Subsequently, the union module fuses the denoised CAMs with region-based CAMs using a weighted combination scheme to obtain more comprehensive and informative maps, which we refer to as fused CAMs. Lastly, the activation map selection module automatically selects the optimal CAM that offers the best interpretation from the pool of fused CAMs. Extensive experiments on ILSVRC2012 and VOC2007 datasets demonstrate UnionCAM's superior performance over state-of-the-art methods. It effectively suppresses background noise, captures complete object regions, and provides intuitive visual explanations. UnionCAM achieves significant improvements in insertion and deletion scores, outperforming the best baseline. UnionCAM makes notable contributions by introducing a novel denoising strategy, adaptive fusion of CAMs, and an automatic selection mechanism. It bridges the gap between CNN performance and interpretability, providing a valuable tool for understanding and trusting CNN-based systems. UnionCAM has the potential to foster responsible deployment of CNNs in real-world applications.

In recent years, the safety issues of high-speed railways have remained severe. The intrusion of personnel or obstacles into the perimeter has often occurred in the past, causing derailment or parking, especially in the case of bad weather such as fog, haze, rain, etc. According to previous research, it is difficult for a single sensor to meet the application needs of all scenario, all weather, and all time domains. Due to the complementary advantages of multi-sensor data such as images and point clouds, multi-sensor fusion detection technology for high-speed railway perimeter intrusion is becoming a research hotspot. To the best of our knowledge, there has been no review of research on multi-sensor fusion detection technology for high-speed railway perimeter intrusion. To make up for this deficiency and stimulate future research, this article first analyzes the situation of high-speed railway technical defense measures and summarizes the research status of single sensor detection. Secondly, based on the analysis of typical intrusion scenarios in high-speed railways, we introduce the research status of multi-sensor data fusion detection algorithms and data. Then, we discuss risk assessment of railway safety. Finally, the trends and challenges of multi-sensor fusion detection algorithms in the railway field are discussed. This provides effective theoretical support and technical guidance for high-speed rail perimeter intrusion monitoring.

With the development of deep learning, very high-resolution remote sensing image change detection (VHRCD) methods are becoming more popular. However, the most existing change detection methods are not good at processing edge details and small target detection. To this end, in this paper, InterFormer, a bidirectional interactive framework based on transformer is proposed to find small changes and extract more accurate edge information of the change area. First, we designed an asymmetric Interaction Attention Module (IAM) to identify the edge details for the bi-temporal image. The IAM fully leverages the benefits of self-attention, performing feature fusion during feature extraction. This approach improves edge feature extraction capability and reduces the number of parameters, compared to other transformer methods. Second, we designed a global attention based feature fusion module called GFFM to enhance the detection performance of small targets. The GFFM further improves small target detection ability by augmenting the network’s selectivity to spatial information during feature fusion. The method is applicable to scenarios involving small changes and possesses enhanced edge detection capabilities. Our method outperforms state-of-the-art counterparts on three public benchmarks and has fewer parameters.

In the real‐world scenario, data often have a long‐tailed distribution and training deep neural networks on such an imbalanced dataset has become a great challenge. The main problem caused by a long‐tailed data distribution is that common classes will dominate the training results and achieve a very low accuracy on the rare classes. Recent work focuses on improving the network representation ability to overcome the long‐tailed problem, while it always ignores adapting the network classifier to a long‐tailed case, which will cause the “incompatibility” problem of network representation and network classifier. In this paper, we use knowledge distillation to solve the long‐tailed data distribution problem and fully optimize the network representation and classifier simultaneously. We propose multiexperts knowledge distillation with class‐balanced sampling to jointly learn high‐quality network representation and classifier. Also, a channel activation‐based knowledge distillation method is also proposed to improve the performance further. State‐of‐the‐art performance on several large‐scale long‐tailed classification datasets shows the superior generalization of our method.

In macro investment, an investment decision model is established by using an improved back propagation (BP) artificial neural network (ANN). In this paper, the relations between elements of investment and output of products are determined, and then the optimal distribution of investment is determined by adjusting the distributions rationally. This model can reflect the highly nonlinear mapping relations among each element of investment by using nonlinear utility functions to improve the architecture of artificial neural network, which can be widely applied in investment problems.