Anime hair design is crucial but challenging, as it conveys personality and emotion through stylized geometry and layered structure. In this work, we propose a sketch-guided approach for intuitive control of multimodal diffusion transformers (MMDiT) to generate semantically consistent anime hairstyles. We adopt a wisp-level flowline input integrated with a fine-tuned MMDiT to transfer hairstyles while preserving character identity. We believe that this fine-grained sketch control within the MMDiT framework may offer a promising path for structured anime hair editing.

Learning from noisy labels remains a major challenge in medical image analysis, where annotation demands expert knowledge and substantial inter-observer variability often leads to inconsistent or erroneous labels. Despite extensive research on learning with noisy labels (LNL), the robustness of existing methods in medical imaging has not been systematically assessed. To address this gap, we introduce LNMBench, a comprehensive benchmark for Label Noise in Medical imaging. LNMBench encompasses \textbf{10} representative methods evaluated across 7 datasets, 6 imaging modalities, and 3 noise patterns, establishing a unified and reproducible framework for robustness evaluation under realistic conditions. Comprehensive experiments reveal that the performance of existing LNL methods degrades substantially under high and real-world noise, highlighting the persistent challenges of class imbalance and domain variability in medical data. Motivated by these findings, we further propose a simple yet effective improvement to enhance model robustness under such conditions. The LNMBench codebase is publicly released to facilitate standardized evaluation, promote reproducible research, and provide practical insights for developing noise-resilient algorithms in both research and real-world medical applications.The codebase is publicly available on https://github.com/myyy777/LNMBench.

Compliant mechanisms enable the creation of compact and easy-to-fabricate devices for tangible interaction. This work explores interconnected compliant mechanisms consisting of multiple joints and rigid bodies to transmit and process displacements as signals that result from physical interactions. As these devices are difficult to design due to their vast and complex design space, we developed a graph-based design algorithm and computational tool to help users program and customize such computational functions and procedurally model physical designs. When combined with active materials with actuation and sensing capabilities, these devices can also render and detect haptic interaction. Our design examples demonstrate the tool's capability to respond to relevant HCI concepts, including building modular physical interface toolkits, encrypting tangible interactions, and customizing user augmentation for accessibility. We believe the tool will facilitate the generation of new interfaces with enriched affordance.

Image rendering from line drawings is vital in design and image generation technologies reduce costs, yet professional line drawings demand preserving complex details. Text prompts struggle with accuracy, and image translation struggles with consistency and fine-grained control. We present LineArt, a framework that transfers complex appearance onto detailed design drawings, facilitating design and artistic creation. It generates high-fidelity appearance while preserving structural accuracy by simulating hierarchical visual cognition and integrating human artistic experience to guide the diffusion process. LineArt overcomes the limitations of current methods in terms of difficulty in fine-grained control and style degradation in design drawings. It requires no precise 3D modeling, physical property specifications, or network training, making it more convenient for design tasks. LineArt consists of two stages: a multi-frequency lines fusion module to supplement the input design drawing with detailed structural information and a two-part painting process for Base Layer Shaping and Surface Layer Coloring. We also present a new design drawing dataset, ProLines, for evaluation. The experiments show that LineArt performs better in accuracy, realism, and material precision compared to SOTAs. Project page: https://meaoxixi.github.io/LineArt/.

The control interfaces of virtual reality (VR) control interfaces have a fundamental trade-off between control naturalness and precision. This paper proposes FingerSlider, a novel single-handed interaction system that integrates optical hand tracking with adaptive parameter mapping to control virtual objects through finger movements. Each finger is mapped to a specific control parameter (e.g., the index finger for the X -axis, the middle finger for the Y -axis), transforming finger movements into virtual sliders. FingerSlider enables simultaneous adjustment of multiple object attributes (position, scale, rotation) while maintaining high control precision and reducing cognitive load. The proposed approach dynamically maps thumb-to-finger sliding distances to control parameters through vector projection and normalization, achieving real-time performance with low computational complexity. Comparison experiments demonstrate that FingerSlider significantly outperformed conventional methods, reducing task completion time by 43 % in single-parameter tasks (24.5 s vs. 43.0 s). While the initial four-finger prototype revealed limitations in multi-finger coordination (SUS score: 58.75), the optimized two-finger system achieved high usability (SUS score: 82.81), representing a 40.95 % improvement. These findings suggest potential applications of FingerSlider in various VR tasks, including 3D object manipulation and content browsing.

Numerous recent works have utilized 3D Gaussian Splatting to represent high-fidelity digital avatars. However, none have enabled interactive multilayer Gaussian garments for virtual try-ons without relying on expensive hardware, such as a camera array and/or multiple GPUs. To enable affordable mix-and-match dressing—dressing 3D avatars with realistic and complex combinations of garments—it is crucial to handle the interactions between multiple layers of garments using consumer-level capturing hardware. To address this, we present a novel screenspace layer resolution method combined with physical simulation and Gaussian garments to enable realistic multilayer mix-and-match avatar dressing at interactive rates using low-cost hardware. As an offline process, we capture multiple static garments individually using only a single mobile camera on a static mannequin and then perform a dual reconstruction of Gaussians and simulation mesh. During runtime, these Gaussians are driven by a fast but simple physics simulator, whose output may contain inter-penetrations across garment layers. Our method fixes these in screenspace by rasterizing the simulation mesh from various camera views and culling the Gaussians that are skinned to unseen mesh triangles. We show the effectiveness of our approach by demonstrating mix-and-match dressing results at interactive rates using short-sleeves, long-sleeves, a fur vest, and a singlet. Additionally, we showcase a webcam-based interactive try-on application to further illustrate the capabilities of our system.

Modern text-to-image generative models can create high-quality and impressive images, but require extensive trial-and-error to interpret user intents. To solve this issue, we propose PromptNavi, a visual exploration interface for node-based prompt composition leveraging large language models to enhance the efficiency of text-to-image generation. In contrast to conventional prompting interfaces, PromptNavi allows users to manipulate and combine visual attributes of target images directly to refine outputs iteratively. Our user study confirmed that the results generated using PromptNavi achieved significant improvements in user usability, reduced cognitive load, and superior image quality rated by independent evaluators. It is verified that users achieved better results with less effort across all measured dimensions, including creativity, atmosphere, coherence, and overall impression. We believe PromptNavi may bridge the gap between user intent and generative AI outputs, advancing human-centered generative AI by making generative models accessible to novices with an enhanced user experience. Source codes are available at: https://anonymous.4open.science/r/project-5996/ . • We propose PromptNavi, a node-and-connection interface that renders prompt engineering more transparent, particularly for non-experts. • An LLM-based approach for fine-grained attribute interpolation, streamlining the process of refining prompts while clarifying the relationship between textual elements and generated outputs. • Empirical validation of PromptNavi’s effectiveness, including user studies demonstrating significant improve- ments in user experience and generative quality over existing baseline tools.

Text-to-image models showcase the impressive ability to generate high-quality and diverse images. However, the transition from freehand sketches to complex scene images with multiple objects remains challenging in computer graphics. In this study, we propose a novel sketch-guided scene image generation framework, decomposing the task of scene image generation from sketch inputs into object-level cross-domain generation and scene-level image construction steps. We first employ a pre-trained diffusion model to convert each single object drawing into a separate image, which can infer additional image details while maintaining the sparse sketch structure. To preserve the conceptual fidelity of the foreground during scene generation, we invert the visual features of object images into identity embeddings for scene generation. For scene-level image construction, we generate the latent representation of the scene image using the separated background prompts. Then, we blend the generated foreground objects with the background image guided by the layout of sketch inputs. We infer the scene image on the blended latent representation using a global prompt with the trained identity tokens to ensure the foreground objects’ details remain unchanged while naturally composing the scene image. Through qualitative and quantitative experiments, we demonstrated that the proposed method’s ability surpasses the state-of-the-art approaches for scene image generation from hand-drawn sketches. • A novel scene sketch-to-image generation method based on text-to-image diffusion models. • The method ensures spatial consistency between generated images and input scene sketches. • Foreground and background are generated separately to balance fidelity and seamless blending.

When employing autonomous driving technology in logistics, small autonomous delivery vehicles (aka delivery robots) encounter challenges different from passenger vehicles when interacting with other road users. We conducted an online video survey as a pre-study and found that autonomous delivery vehicles need external human-machine interfaces (eHMIs) to ask for help due to their small size and functional limitations. Inspired by everyday human communication, we chose arms as eHMI to show their request through limb motion and gesture. We held an in-house workshop to identify the arm’s requirements for designing a specific arm with shrink-ability (conspicuous when delivering messages but not affect traffic at other times). We prototyped a small delivery robot with a shrinkable arm and filmed the experiment videos. We conducted two studies (a video-based and a 360-degree-photo VR-based) with 18 participants. We demonstrated that arm-on-delivery robots can increase interaction efficiency by drawing more attention and communicating specific information.

Human-centered AI aims to bridge the gap between machine decision-making and human understanding. However, even for classification tasks where deep neural networks have achieved superb performance, there are currently few methods that link humans and AI well, especially on domain-specific tasks. In this paper, we propose SpaceEditing, a 2D spatial layout tool that enables human users to interact with the latent space of deep neural networks. During the interaction process, the tool’s algorithm automatically processes user actions, providing feedback to the network and leveraging triplet loss to effectively learn from user-modified information. We evaluate SpaceEditing with three case studies: (1) an archaeology researcher uses a bronze dataset; (2) a deep learning researcher uses a garbage classification dataset; (3) six deep learning beginners use a head pose dataset. The experimental results demonstrate the effectiveness of our tool in integrating human knowledge and improving network performance.

The dynamic effects of smoke are impressive in illustration design, but it is a troublesome and challenging issue for inexpert users to design smoke effects without domain knowledge of fluid simulations. In this work, we propose DualSmoke, a two-stage global-to-local generation framework for interactive smoke illustration design. In the global stage, the proposed approach utilizes fluid patterns to generate Lagrangian coherent structures from the user’s hand-drawn sketches. In the local stage, detailed flow patterns are obtained from the generated coherent structure. Finally, we apply a guiding force field to the smoke simulator to produce the desired smoke illustration. To construct the training dataset, DualSmoke generates flow patterns using finite-time Lyapunov exponents of the velocity fields. The synthetic sketch data are generated from the flow patterns by skeleton extraction. Our user study verifies that the proposed design interface can provide various smoke illustration designs with good user usability. Our code is available at https://githubcom/shasph/DualSmoke.

Attention guidance is used to address dataset bias in deep learning, where the model relies on incorrect features to make decisions. Focusing on image classification tasks, we propose an efficient human-in-the-loop system to interactively direct the attention of classifiers to regions specified by users, thereby reducing the effect of co-occurrence bias and improving the transferability and interpretability of a deep neural network (DNN). Previous approaches for attention guidance require the preparation of pixel-level annotations and are not designed as interactive systems. We herein present a new interactive method that allows users to annotate images via simple clicks. Additionally, we identify a novel active learning strategy that can significantly reduce the number of annotations. We conduct both numerical evaluations and a user study to evaluate the proposed system using multiple datasets. Compared with the existing non-active-learning approach, which typically relies on considerable amounts of polygon-based segmentation masks to fine-tune or train the DNNs, our system can obtain fine-tuned networks on biased datasets in a more time- and cost-efficient manner and offers a more user-friendly experience. Our experimental results show that the proposed system is efficient, reasonable, and reliable. Our code is publicly available at https://github.com/ultratykis/Guiding-DNNs-Attention.

In this study, we propose an efficient city-generation method based on user sketches that combine a deep generative model with the procedural modeling approach. The proposed framework adopts the deep learning-based network of conditional generative adversarial networks. For the data training, we randomly generated three-dimensional cities from perlin noise. The contours of the cities were extracted by the holistically-nested edge detection approach. The proposed method is a deep learning model that uses paired data of cities generated by a procedural model, along with the corresponding hand-drawn style sketches.

Recently, haptic gloves have been extensively explored for various practical applications, such as manipulation learning. Previous glove devices have different force-driven systems, such as shape memory alloys, servo motors and pneumatic actuators; however, these proposed devices may have difficulty in fast finger movement, easy reproduction, and safety issues. In this study, we propose MagGlove, a novel haptic glove with a movable magnet mechanism that has a linear motor, to solve these issues. The proposed MagGlove device is a compact system on the back of the wearer’s hand with high responsiveness, ease of use, and good safety. The proposed device is adaptive with the modification of the magnitude of the current flowing through the coil. Based on our evaluation study, it is verified that the proposed device can achieve finger motion in the given tasks. Therefore, MagGlove can provide flexible support tailored to the wearers’ learning levels in manipulation learning tasks.

Drawing the facial features of anime characters at different ages is challenging in the creation process. Since characters’ facial features at different ages have obvious differences, it is difficult, especially for novices to accurately illustrate the age features of anime characters. Conventional data-driven drawing interfaces for anime characters focus on the visual features of hair, emotion, and coloring but fail to provide age-specific drawing guidance. To address this gap, we propose AgeFace, an interactive drawing interface that assists users in creating facial features with age-specific features based on user input strokes. We evaluated the usability of AgeFace by a user experience experiment and a comparison experiment with baseline approaches. The results verified that AgeFace could achieve better performance in usability and better support in the creative process than baseline systems.

This research focuses on interpreting the Eshelby twist behavior into kinetic façade design, which is a continuous crystallographic twist generated by the torsional force of an axial screw dislocation in a one-dimensional nanowire structure. The mechanism can be adjusted to the kinetic façade, changing dynamically rather than being static or fixed and allowing movement to occur on a building's surface with the potential to reduce sunlight exposure. Natural light is essential for making human work in building spaces more productive. The solution to the issue of inappropriate natural light in indoor spaces is a kinetic façade, which is self-adjustable according to sunlight intensity. The simulation results before and after installing the façade show the sunlight exposure value were 2030 kWh/m2and 300 kWh/m2per year, respectively. 2030 kWh/m2, this value is too excessive for direct sunlight for users living in the space. As a result, the sunlight exposure value after installation is acceptable for users; it is verified that the proposed design can strongly decrease sunlight exposure. These findings indicate that the Eshelby twist is the principal phenomenon facilitating façade movement.

Scene graphs have been widely used in various visual applications, such as image retrieval and generation. However, it is time consuming and challenging to draw scene graphs, especially for images with complex scenes. To resolve this issue, we propose an interactive editing interface for scene graph representation in which users can draw scene graphs simply and conveniently. We provide alternatives for frequently used objects, attributes, and relationships in graph drawing. The proposed function design can greatly reduce the drawing time cost and improve drawing quality. We conducted a user study and confirmed that the proposed interface can help users draw the desired scene graph, as well as providing a good user experience.

Deep neural networks (DNNs) have achieved excellent results currently in classification, while they may still suffer from ambiguous images which are similar across classes. By contrast, humans have a relatively good ability to distinguish these categories of images. Therefore, we propose a human-in-the-loop solution to assist the network to better classify the images by leveraging human knowledge. To achieve this, we project the high-dimensional latent space trained by the network onto a two-dimensional workspace. The users can interactively modify the projected coordinates of inputs on the workspace using our designed tools, then the modified information will be fed back to the network to fine-tune it, which in turn affects the network's classification results, thereby improving the accuracy of network classification.

Special skills are required in portrait painting, such as imagining geometric structures and facial detail for final portrait designs. This makes it a difficult task for users, especially novices without prior artistic training, to draw freehand portraits with high-quality details. In this paper, we propose dualFace, a portrait drawing interface to assist users with different levels of drawing skills to complete recognizable and authentic face sketches. Inspired by traditional artist workflows for portrait drawing, dualFace gives two-stages of drawing assistance to provide global and local visual guidance. The former helps users draw contour lines for portraits (i.e., geometric structure), and the latter helps users draw details of facial parts, which conform to the user-drawn contour lines. In the global guidance stage, the user draws several contour lines, and dualFace then searches for several relevant images from an internal database and displays the suggested face contour lines on the background of the canvas. In the local guidance stage, we synthesize detailed portrait images with a deep generative model from user-drawn contour lines, and then use the synthesized results as detailed drawing guidance. We conducted a user study to verify the effectiveness of dualFace, which confirms that dualFace significantly helps users to produce a detailed portrait sketch.

Abstract In this paper, the problem of clustering data points that lie near or on a union of independent low‐dimensional subspaces is addressed. To this end, the popular spectral clustering‐based algorithms usually follow a two‐stage strategy that initially builds an affinity matrix and then applies spectral clustering. However, an inappropriate affinity matrix that does not sufficiently connect data points lying on the same subspace will easily lead to the issue of over‐segmentation. To alleviate this issue, building the affinity matrix based on subspace hypotheses generated by an iterative sampling operation according to the Random Cluster Model under the framework of energy minimisation is proposed. Specifically, each hypothesis is generated from a large number of data points by sampling a component in a K ‐nearest neighbour graph. Extensive experiments on synthetic data and real‐world datasets show that the proposed method can improve the connectivity of the affinity matrix and provide competitive results against state‐of‐the‐art methods.

In this paper, we propose a user interface that enables users to intuitively retrieve relevant motions from a database and edit them by drawing motion trajectories on the screen. This system consists of two-stage operations to provide global-level and local-level motion editing: a global stage that enables users to design the body movement in virtual space roughly, and a local stage that enables users to design detailed movements such as limbs movement. We verified the proposed system with character animation editing with both global and local stages.

The cover image is based on the Special Issue Paper Sketch2VF: Sketch-based Flow Design with Conditional Generative Adversarial Network by Xie Haoran et al., https://doi.org/10.1002/cav.1889.

Abstract We present an interactive user interface to support sketch‐based fluid design with a perceptual understanding of human sketches. In particular, the proposed system generates a 2D fluid animation from hand‐drawn sketches. The proposed system utilizes a conditional generative adversarial network model to generate stationary velocity fields from a sketch input. The network model is trained with hand‐drawn strokes and corresponding 2D velocity fields. On the basis of the generated velocity field, the system calculates fluid dynamics using a semi‐Lagrangian method. We ran a user study of the proposed system and confirmed that the proposed interface is effective for a 2D fluid design and that the system achieves good results based on user input.

In this article, we introduce an efficient and versatile numerical aerodynamics model for general three-dimensional geometry shapes in potential flow. The proposed model has low computational cost and achieves an accuracy of moderate fidelity for the aerodynamic loads for a given glider shape. In the geometry preprocessing steps of our model, lifting-wing surfaces are recognized, and wake panels are generated automatically along the trailing edges. The proposed aerodynamics model improves the potential theory-based panel method. Furthermore, a new quadratic expression for aerodynamic forces and moments is proposed. It consists of geometry-dependent aerodynamic coefficient matrices and has a continuous representation for the drag/lift-force coefficients. Our model enables natural and real-time aerodynamics simulations combined with general rigid-body simulators for interactive animation. We also present a design system for original gliders. It uses an assembly-based modeling interface and achieves interactive feedback by leveraging the partwise precomputation enabled by our method. We illustrate that one can easily design various flyable gliders using our system.

This paper proposes a pattern-guided framework for immersed rigid body simulations involving unsteady dynamics of a fully immersed or submerged rigid body in a still flow. Instead of the heavy computation of fluid-body coupling simulations, a novel framework considering different flow effects from the surrounding flow is constructed by parameter estimation of force coefficients. We distinguish the flow effects of the inertial, viscous and turbulent effects to the rigid body. It is difficult to clarify the force coefficients of viscous effect in real flow. In this paper we define the control parameters of viscous forces in rigid body simulator, and propose a energy optimization strategy for determining the time series of control parameters. This strategy is built upon a motion graph of motion patterns and the turbulent kinetic energy. The proposed approach achieves efficient and realistic immersed rigid body simulation results, and these results are relevant to the real-time animations of body-vorticity coupling.

We present in this paper a framework that generates free fall motions for the object within a still fluid. We introduce a new motion synthesis approach where six characteristic motion prototypes of free fall are defined and synthesized, and then the motion trajectory is specified form free fall motion graph. We automatically create motion sequences using trajectory search tree and pre-computed trajectory database. The proposed approach can produce realistic and controllable free fall motion that could be applied in many different applications, including virtual reality, game and other entertainment productions.