Embodied navigation is a fundamental capability that enables embodied agents to effectively interact with the physical world in various complex environments. However, a significant gap remains between current embodied navigation tasks and real-world requirements, as existing methods often struggle to integrate high-level human instructions with spatial understanding. To address this gap, we propose a new task of embodied navigation called spatial navigation, which encompasses two key components: spatial object navigation (SpON) for object-specific guidance and spatial area navigation (SpAN) for navigating to designated areas. Specifically, SpON guides agents to specific objects by leveraging spatial relationships and contextual understanding, while SpAN focuses on navigating to defined areas within complex environments. Together, these components significantly enhance agents’ navigation capabilities, enabling more effective interactions in real-world scenarios. To support this task, we have generated a spatial navigation dataset consisting of 10K trajectories within the simulator. This dataset includes high-level human instructions, detailed observations, and corresponding navigation actions, providing a comprehensive resource to enhance agent training and performance. Building on the spatial navigation dataset, we introduce SpNav, a hierarchical navigation framework. Specifically, SpNav employs vision-language model (VLM) to interpret high-level human instructions and accurately identify goal objects or areas within the observation range, achieving precise point-to-point navigation using a map and enhancing the agent’s ability to oper- ate effectively in complex environments by bridging the gap between perception and action. Extensive experiments show that SpNav achieves state-of-the-art (SOTA) performance in spatial navigation tasks across both simulated and real-world environments, validating the effectiveness of our method.

Embodied navigation and manipulation are fundamental capabilities for embodied agents operating in physical environments. A key challenge in this process is understanding the spatial context and the affordances of the environment, which involves recognizing how objects can be interacted with (object affordance) and identifying suitable locations for movement and object placement (free space affordance). While Vision-Language Models (VLMs) have shown promise in high-level task planning, their ability to translate reasoning into precise executable actions remains limited, particularly in image-based spatial understanding and precise affordance localization-a critical gap in image processing for robotics. To bridge this gap, we propose EspA, a novel image-to-keypoint model that leverages spatial-aware affordance learning to predict actionable affordances directly from 2D image inputs. Built on a hierarchical vision-language architecture, EspA jointly reasons about object affordances and free space affordances, enabling pixel-level localization of both types of interactions. Crucially, EspA translates language instructions into precise 2D affordance keypoints from observed images, which are then projected into 3D actionable coordinates using depth information. To support this unified affordance reasoning, we introduce the Embodied Spatial Affordance (ESA) dataset, which captures both object-centric interactions and free space contexts. By jointly modeling these affordances in a shared representation space, EspA overcomes the limitations of prior works that treat them independently. The dataset's fine-grained annotations enable our model to learn the intricate relationship between object functionality and spatial feasibility, significantly enhancing the spatial understanding in embodied tasks. Extensive experimental results demonstrate that EspA outperforms existing state-of-the-art Vision-Language Models (VLMs), both open-source and closed-source, in object and free space affordance prediction. Furthermore, it exhibits superior performance in real-world embodied navigation and manipulation experiments. Our work advances the field of image-based spatial reasoning by providing a scalable solution for translating high-level instructions into low-level actionable affordances. We believe this work paves the way for more robust and versatile embodied agents capable of effectively interacting with complex environments. The dataset, benchmark, and evaluation code will be publicly available to facilitate future research. Project website: https://embodied-spatial-affordance.github.io/.

Integrating large language models (LLMs) into autonomous driving has attracted significant attention with the hope of improving generalization and explainability. However, existing methods often focus on either driving or vision-language understanding but achieving both high driving performance and extensive language understanding remains challenging. In addition, the dominant approach to tackle vision-language understanding is using visual question answering. However, for autonomous driving, this is only useful if it is aligned with the action space. Otherwise, the model’s answers could be inconsistent with its behavior. Therefore, we propose a model that can handle three different tasks: (1) closed-loop driving, (2) vision-language understanding, and (3) language-action alignment. Our model SimLingo is based on a vision language model (VLM) and works using only camera, excluding expensive sensors like LiDAR. SimLingo obtains state-of-the-art performance on the widely used CARLA simulator on the Bench2Drive benchmark and is the winning entry at the CARLA challenge 2024. Additionally, we achieve strong results in a wide variety of language-related tasks while maintaining high driving performance. Project page: https://katrinrenz.de/simlingo

This paper reviews the VCIP 2025 Grand Challenge on Live Broadcasting Video Quality Assessment. The competition aims to foster innovation in both subjective and objective VQA techniques tailored to live broadcasting videos, addressing the unique challenges posed by live streaming impairments while emphasizing the evaluation of QoE. The grand challenge used live broadcasting database LBVD which consists of 1013 videos focusing on distortion in live broadcasting videos. The competition had 14 participants and 5 teams submitted valid solutions for the final testing phase. The proposed solutions have shown significant progress in areas such as combining traditional feature engineering with deep learning models, achieved state-of-the-art performances for LBVD. Team ATHENA-Live-QoE and Team HZX Force tied for the first position. The dataset can be found at https://github.com/cpf0079/LBVD.

Large Language Models (LLMs) have shown promise in the autonomous driving sector, particularly in generalization and interpretability. We introduce a unique objectlevel multimodal LLM architecture that merges vectorized numeric modalities with a pre-trained LLM to improve context understanding in driving situations. We also present a new dataset of 160k QA pairs derived from 10k driving scenarios, paired with high quality control commands collected with RL agent and question answer pairs generated by teacher LLM (GPT-3.5). A distinct pretraining strategy is devised to align numeric vector modalities with static LLM representations using vector captioning language data. We also introduce an evaluation metric for Driving QA and demonstrate our LLM-driver’s proficiency in interpreting driving scenarios, answering questions, and decision-making. Our findings highlight the potential of LLM-based driving action generation in comparison to traditional behavioral cloning. We make our benchmark, datasets, and model available1for further exploration.

With the world-wide development of 2019 novel coronavirus, although WHO has officially announced the disease as COVID-19, one controversial term - "Chinese Virus" is still being used by a great number of people. In the meantime, global online media coverage about COVID-19-related racial attacks increases steadily, most of which are anti-Chinese or anti-Asian. As this pandemic becomes increasingly severe, more people start to talk about it on social media platforms such as Twitter. When they refer to COVID-19, there are mainly two ways: using controversial terms like "Chinese Virus" or "Wuhan Virus", or using non-controversial terms like "Coronavirus". In this article, we attempt to characterize the Twitter users who use controversial terms and those who use non-controversial terms. We use the Tweepy API to retrieve 17 million related tweets and the information of their authors. We find the significant differences between these two groups of Twitter users across their demographics, user-level features like the number of followers, political following status, as well as their geo-locations. Moreover, we apply classification models to predict Twitter users who are more likely to use controversial terms. To our best knowledge, this is the first large-scale social media-based study to characterize users with respect to their usage of controversial terms during a major crisis.

Abstract Mixed reality (MR) is a powerful interactive technology for new types of user experience. We present a semantic‐based interactive MR framework that is beyond current geometry‐based approaches, offering a step change in generating high‐level context‐aware interactions. Our key insight is that by building semantic understanding in MR, we can develop a system that not only greatly enhances user experience through object‐specific behaviours, but also it paves the way for solving complex interaction design challenges. In this paper, our proposed framework generates semantic properties of the real‐world environment through a dense scene reconstruction and deep image understanding scheme. We demonstrate our approach by developing a material‐aware prototype system for context‐aware physical interactions between the real and virtual objects. Quantitative and qualitative evaluation results show that the framework delivers accurate and consistent semantic information in an interactive MR environment, providing effective real‐time semantic‐level interactions.

Surgical smoke removal algorithms can improve the quality of intra-operative imaging and reduce hazards in image-guided surgery, a highly desirable post-process for many clinical applications. These algorithms also enable effective computer vision tasks for future robotic surgery. In this article, we present a new unsupervised learning framework for high-quality pixel-wise smoke detection and removal. One of the well recognized grand challenges in using convolutional neural networks (CNNs) for medical image processing is to obtain intra-operative medical imaging datasets for network training and validation, but availability and quality of these datasets are scarce. Our novel training framework does not require ground-truth image pairs. Instead, it learns purely from computer-generated simulation images. This approach opens up new avenues and bridges a substantial gap between conventional non-learning based methods and which requiring prior knowledge gained from extensive training datasets. Inspired by the Generative Adversarial Network (GAN), we have developed a novel generative-collaborative learning scheme that decomposes the de-smoke process into two separate tasks: smoke detection and smoke removal. The detection network is used as prior knowledge, and also as a loss function to maximize its support for training of the smoke removal network. Quantitative and qualitative studies show that the proposed training framework outperforms the state-of-the-art de-smoking approaches including the latest GAN framework (such as PIX2PIX). Although trained on synthetic images, experimental results on clinical images have proved the effectiveness of the proposed network for detecting and removing surgical smoke on both simulated and real-world laparoscopic images.

This paper proposes a stable and accurate object registration pipeline for markerless augmented reality applications. We present two novel algorithms for object recognition and matching to improve the registration accuracy from model to scene transformation via point cloud fusion. Whilst the first algorithm effectively deals with simple scenes with few object occlusions, the second algorithm handles cluttered scenes with partial occlusions for robust real-time object recognition and matching. The computational framework includes a locally supported Gaussian weight function to enable repeatable detection of 3D descriptors. We apply a bilateral filtering and outlier removal to preserve edges of point cloud and remove some interference points in order to increase matching accuracy. Extensive experiments have been carried to compare the proposed algorithms with four most used methods. Results show improved performance of the algorithms in terms of computational speed, camera tracking and object matching errors in semi-cluttered and partial-occluded scenes.

Markerless Augmented Reality (AR) registration using the standard Homography matrix is unstable, and for image-based registration it has very low accuracy. In this paper, we present a new method to improve the stability and the accuracy of marker-less registration in AR. Based on the Visual Simultaneous Localization and Mapping (V-SLAM) framework, our method adds a three-dimensional dense cloud processing step to the state-of-the-art ORB-SLAM in order to deal with mainly the point cloud fusion and the object recognition. Our algorithm for the object recognition process acts as a stabilizer to improve the registration accuracy during the model to the scene transformation process. This has been achieved by integrating the Hough voting algorithm with the Iterative Closest Points(ICP) method. Our proposed AR framework also further increases the registration accuracy with the use of integrated camera poses on the registration of virtual objects. Our experiments show that the proposed method not only accelerates the speed of camera tracking with a standard SLAM system, but also effectively identifies objects and improves the stability of markerless augmented reality applications.

Mixed Reality (MR) is of increasing interest within technology-driven modern medicine but is not yet used in everyday practice. This situation is changing rapidly, however, and this paper explores the emergence of MR technology and the importance of its utility within medical applications. A classification of medical MR has been obtained by applying an unbiased text mining method to a database of 1,403 relevant research papers published over the last two decades. The classification results reveal a taxonomy for the development of medical MR research during this period as well as suggesting future trends. We then use the classification to analyse the technology and applications developed in the last five years. Our objective is to aid researchers to focus on the areas where technology advancements in medical MR are most needed, as well as providing medical practitioners with a useful source of reference.