6-DoF Grasp Detection Method Based on Vision Language Guidance

The increasing integration of Visual Language Models (VLMs) into visualization systems demands a comprehensive understanding of their visual interpretation capabilities and constraints. While existing research has examined individual models, systematic comparisons of VLMs' visualization literacy remain unexplored. We bridge this gap through a rigorous, first‐of‐its‐kind evaluation of four leading VLMs (GPT‐4, Claude, Gemini, and Llama) using standardized assessments: the Visualization Literacy Assessment Test (VLAT) and Critical Thinking Assessment for Literacy in Visualizations (CALVI). Our methodology uniquely combines randomized trials with structured prompting techniques to control for order effects and response variability ‐ a critical consideration overlooked in many VLM evaluations. Our analysis reveals that while specific models demonstrate competence in basic chart interpretation (Claude achieving 67.9% accuracy on VLAT), all models exhibit substantial difficulties in identifying misleading visualization elements (maximum 30.0% accuracy on CALVI). We uncover distinct performance patterns: strong capabilities in interpreting conventional charts like line charts (76‐96% accuracy) and detecting hierarchical structures (80‐100% accuracy), but consistent difficulties with data‐dense visualizations involving multiple encodings (bubble charts: 18.6‐61.4%) and anomaly detection (25‐30% accuracy). Significantly, we observe distinct uncertainty management behavior across models, with Gemini displaying heightened caution (22.5% question omission) compared to others (7‐8%). These findings provide crucial insights for the visualization community by establishing reliable VLM evaluation benchmarks, identifying areas where current models fall short, and highlighting the need for targeted improvements in VLM architectures for visualization tasks. To promote reproducibility, encourage further research, and facilitate benchmarking of future VLMs, our complete evaluation framework, including code, prompts, and analysis scripts, is available at https://github.com/washuvis/VisLit‐VLM‐Eval.

With the development of deep learning technology, the facial images generated by deepfake technology have reached a level of authenticity that is difficult to distinguish, posing a serious threat to personal privacy and data security. Therefore, it is of great significance to develop efficient and reliable deepfake detection technology. In recent years, Visual Language Models (VLM) have been applied to deepfake detection tasks due to their powerful multimodal understanding capabilities. However, the existing VLM have not been specifically optimized for deepfake detection tasks. When directly applied to this task, there are problems such as insufficient model accuracy and insufficient feature extraction, especially when dealing with complex forgery scenes. In response to these challenges, this paper proposes an innovative deepfake face detection method based on VLM and component-specific prompt tuning. We transform the deepfake detection task into a Visual Question Answering (VQA) task, making full use of the multimodal understanding capabilities of VLM and the flexibility of prompt tuning technology. This method uses a local prompt strategy to customize specific prompt questions for key facial components such as eyes, nose, and mouth, guiding the model to focus on the local features of these areas, thereby accurately capturing forgery traces. In addition, we introduced a feature extraction module Q-Former based on instructions, which can flexibly adjust the focus area of visual features according to prompts, significantly improving the model’s perception of locally forged features. By fusing these local features extracted by Q-Former and combining them with the language model to judge the authenticity of the overall face image, we can finally generate accurate prediction results. A large number of experimental results show that our method is significantly better than existing technologies in terms of detection accuracy and robustness.

Automated task planning traditionally relies on manually generated domain models, creating bottlenecks in scalability and requiring extensive domain expertise. This paper presents a novel framework to automate the process of generating Planning Domain Definition Language (PDDL) domains and problem files by integrating Web Ontology Language (OWL) ontologies with Visual Language Models (VLMs). Our approach leverages the rich semantic structure of OWL ontologies to systematically define domain classes, predicates, and actions, while VLMs ground abstract ontological concepts in concrete visual observations—automating the generation of instance‑specific planning problems. The proposed framework transforms ontological knowledge into PDDL domain files through a mapping algorithm that preserves semantic relationships and logical constraints. The VLM performs visual scene analysis to identify relevant objects, attributes, and spatial configurations for generating initial states, while natural language instructions are used to derive goal states. We evaluate the framework across multiple planning domains, demonstrating that it generates syntactically correct and semantically coherent PDDL domain and problem files directly from OWL ontologies, camera images, and natural language inputs. The resulting files are comparable in quality to those manually generated by planning experts and outperform previous automated systems in terms of semantic fidelity and adaptability.

Large Language Models (LLMs) and Visual Language Models (VLMs) are attracting increasing interest due to their improving performance and applications across various domains and tasks. However, LLMs and VLMs can produce erroneous results, especially when a deep understanding of the problem domain is required. For instance, when planning and perception are needed simultaneously, these models often struggle because of difficulties in merging multi-modal information. To address this issue, fine-tuned models are typically employed and trained on specialized data structures representing the environment. This approach has limited effectiveness, as it can overly complicate the context for processing. In this paper, we propose a multi-agent architecture for embodied task planning that operates without the need for specific data structures as input. Instead, it uses a single image of the environment, handling free-form domains by leveraging commonsense knowledge. We also introduce a novel, fully automatic evaluation procedure, PG2S, designed to better assess the quality of a plan. We validated our approach using the widely recognized ALFRED dataset, comparing PG2S to the existing KAS metric to further evaluate the quality of the generated plans.

Humans can infer the affordance of objects by extracting related contextual preconditions for each scenario. For example, upon seeing an image of a broken cup, we can infer that this precondition prevents the cup from being used for drinking. Reasoning with preconditions of commonsense is studied in NLP where the model explicitly gets the contextual precondition. However, it is unclear if SOTA visual language models (VLMs) can extract such preconditions and infer the affordance of objects with them. In this work, we introduce the task of preconditioned visual language inference and rationalization (PVLIR). We propose a learning resource based on three strategies to retrieve weak supervision signals for the task and develop a human-verified test set for evaluation. Our results reveal the shortcomings of SOTA VLM models in the task and draw a road map to address the challenges ahead in improving them.

A Real-World Size Organization of Object Responses in Occipitotemporal Cortex

A class of visual programming languages whose members share a common group of underlying principles is identified. In this class (the demand-driven temporal-assignment visual language model), the elimination of certain duplicate computations is a natural by-product of the model. The potential time and space complexity characteristics for visual languages based on this model are discussed, and a method that makes use of these characteristics is presented. >

The rapid advancement of Visual Language Models (VLMs) has enabled sophisticated analysis of visual content, leading to concerns about the inference of sensitive user attributes and subsequent privacy risks. While technical capabilities of VLMs are increasingly studied, users’ understanding, perceptions, and reactions to these inferences remain less explored, especially concerning videos uploaded on the social media. This paper addresses this gap through a semi-structured interview (N=17), investigating user perspectives on VLM-driven sensitive attribute inference from their visual data. Findings reveal that users perceive VLMs as capable of inferring a range of attributes, including location, demographics, and socioeconomic indicators, often with unsettling accuracy. Key concerns include unauthorized identification, misuse of personal information, pervasive surveillance, and harm from inaccurate inferences. Participants reported employing various mitigation strategies, though with skepticism about their ultimate effectiveness against advanced AI. Users also articulate clear expectations for platforms and regulators, emphasizing the need for enhanced transparency, user control, and proactive privacy safeguards. These insights are crucial for guiding the development of responsible AI systems, effective privacy-enhancing technologies, and informed policymaking that aligns with user expectations and societal values.

The integration of Large Language Models (LLMs), particularly Visual Language Models (VLMs), into robotics promises more intuitive human–robot interactions; however, challenges remain in efficiently translating high-level commands into precise physical actions. This paper presents a novel architecture for vision-based object manipulation that leverages a VLM’s reasoning capabilities while incorporating symmetry principles to enhance operational efficiency. Implemented on a Yahboom DOFBOT educational robot with a Jetson Nano platform, our system introduces a prompt-based framework that uniquely embeds symmetry-related cues to streamline feature extraction and object detection from visual data. This methodology, which utilizes few-shot learning, enables the VLM to generate more accurate and contextually relevant commands for manipulation tasks by efficiently interpreting the symmetric and asymmetric features of objects. The experimental results in controlled scenarios demonstrate that our symmetry-informed approach significantly improves the robot’s interaction efficiency and decision-making accuracy compared to generic prompting strategies. This work contributes a robust method for integrating fundamental vision principles into modern generative AI workflows for robotics. Furthermore, its implementation on an accessible educational platform shows its potential to simplify complex robotics concepts for engineering education and research.

Often the most effective diagrams are those which are very simple. However there is a strong tendency, particularly prevalent in visual formal modelling and specification languages, to take a diagrammatic language which at core is very simple, and then add many extensions and features to make it more expressive - often making it so expressive that the diagrams produced in the language are no longer readable; or at least, the diagrams are no longer obviously a more effective form of representation than a text-based one. Both the design of effective visual formal modelling and specification languages, and the effective formalisation of (the semantics of) such visual languages requires the unification of results from visual language theory, cognitive science, empirical psychology and graphic design. Integrating results from such diverse fields is a non-trivial task, which may be approached through a decomposition of the study of issues of effectiveness in diagrammatic languages according to analogous understandings of (written and spoken) natural languages.

In recent years, with the development of deep learning, machine translation using neural network has gradually become the mainstream method in industry and academia. The existing Chinese-English machine translation models generally adopt the deep neural network architecture based on attention mechanism. However, it is still a challenging problem to model short and long sequences simultaneously. Therefore, a bidirectional LSTM model integrating attention mechanism is proposed. Firstly, by using the word vector as the input data of the translation model, the linguistic symbols used in the translation process are mathematized. Secondly, two attention mechanisms are designed: local attention mechanism and global attention mechanism. The local attention mechanism is mainly used to learn which words or phrases in the input sequence are more important for modeling, while the global attention mechanism is used to learn which layer of expression vector in the input sequence is more critical. Bidirectional LSTM can better fuse the feature information in the input sequence, while bidirectional LSTM with attention mechanism can simultaneously model short and long sequences. The experimental results show that compared with many existing translation models, the bidirectional LSTM model with attention mechanism can effectively improve the quality of machine translation.

The pursuit of Artificial Intelligence (AI) that emulates human cognitive processes is a cornerstone of ethical AI development, ensuring that emerging technologies can seamlessly integrate into societal frameworks requiring nuanced understanding and decision-making. Zero-Shot Instance Retrieval (ZSIR) stands at the forefront of this endeavour, potentially providing a robust platform for AI systems, particularly large visual language models, to demonstrate and refine cognition-aligned learning without the need for direct experience. In this paper, we critically evaluate current cognition alignment methodologies within traditional zero-shot learning paradigms using visual attributes and word embedding generated by large AI models. We propose a unified similarity function that quantifies the cognitive alignment level, bridging the gap between AI processes and human-like understanding. Through extensive experimentation, our findings illustrate that this similarity function can effectively mirror the visual–semantic gap, steering the model towards enhanced performance in Zero-Shot Instance Retrieval. Our models achieve state-of-the-art performance on both the SUN (92.8% and 82.2%) and CUB datasets (59.92% and 48.82%) for bi-directional image-attribute retrieval accuracy. This work not only benchmarks the cognition alignment of AI but also sets a new precedent for the development of visual language models attuned to the complexities of human cognition.

In this letter, we present a transformer-based architecture, namely TF-Grasp, for robotic grasp detection. The developed TF-Grasp framework has two elaborate designs making it well suitable for visual grasping tasks. The first key design is that we adopt the local window attention to capture local contextual information and detailed features of graspable objects. Then, we apply the cross window attention to model the long-term dependencies between distant pixels. Object knowledge, environmental configuration, and relationships between different visual entities are aggregated for subsequent grasp detection. The second key design is that we build a hierarchical encoder-decoder architecture with skip-connections, delivering shallow features from the encoder to decoder to enable a multi-scale feature fusion. Due to the powerful attention mechanism, TF-Grasp can simultaneously obtain the local information (i.e., the contours of objects), and model long-term connections such as the relationships between distinct visual concepts in clutter. Extensive computational experiments demonstrate that TF-Grasp achieves competitive results versus state-of-art grasping convolutional models and attains a higher accuracy of <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$97.99 \%$</tex-math></inline-formula> and 94.6% on Cornell and Jacquard grasping datasets, respectively. Real-world experiments using a 7DoF Franka Emika Panda robot also demonstrate its capability of grasping unseen objects in a variety of scenarios. The code is available at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/WangShaoSUN/grasp-transformer</uri> .

With the introduction of deep learning, a significant amount of research has been conducted in the field of computer vision in the past decade. In particular, research on object detection (OD) continues to progress rapidly. However, despite these advances, some limitations need to be overcome to enable real-world applications of deep learning-based OD models. One such limitation is inaccurate OD when image quality is poor or a target object is small. The performance degradation phenomenon for small objects is similar to the fundamental limitations of an OD model, such as the constraint of the receptive field, which is a difficult problem to solve using only an OD model. Therefore, OD performance can be hindered by low image quality or small target objects. To address this issue, this study investigates the compatibility of super-resolution (SR) and OD techniques to improve detection, particularly for small objects. We analyze the combination of SR and OD models, classifying them based on architectural characteristics. The experimental results show a substantial improvement when integrating OD detectors with SR models. Overall, it was demonstrated that, when the evaluation metrics (PSNR, SSIM) of the SR models are high, the performance in OD is correspondingly high as well. Especially, evaluations on the MS COCO dataset reveal that the enhancement rate for small objects is 9.4% higher compared to all objects. This work provides an analysis of SR and OD model compatibility, demonstrating the potential benefits of their synergistic combination. The experimental code can be found on our GitHub repository.

Visual Large Language Models in Radiology: A Systematic Multimodel Evaluation of Diagnostic Accuracy and Hallucinations