Complete List of Posters

A dynamic 3D model of human swallowing was created from multi-slice CT images (3D + time) by using Blender for geometry creation and Smoothed Particle Hydrodynamics for fluid simulation.
Invited to the First Round of the Student Research Competition

A new quantization scheme for the second layer of two-layer HDR encoding methods that better preserves residual information that could not be encoded in the first layer.

An integrated robotic system that makes tedious procedures, such as multiview photometric stereo scanning, possible in practical processing times and allows generation of accurate high-resolution 3D models of anaesthetised lizards for analysis of colour time-evolution.

This system scans objects and displays them as stereoscopic 3DCG. Putting objects on the screen, the system scans the objects and generates stereoscopic 3DCG with motion parallax. Users can observe them from any position in all directions and also touch them with their hands.

There has been a lot of research on water-drop models, and on melting ice and icicles. This model can deal with the phase-transition process from start to end, from water to ice, by extending the SPH method.

In recent years, various tablet devices have become widely used. However, users cannot feel a realistic tactile impression when they press the virtual button key on a tablet device. This project proposes an unconstrained realistic tactile rendering with tablet device based on time-series haptic sensing with bilateral control.

A CPU-GPU hybrid architecture with an extended asynchronous bounding-volume hierarchy construction scheme for ray tracing dynamic scenes.

A novel method that can deblur facial images using a patch-based texture synthesis with a very small image database.

Methods that can add invisible information to an actual object and then visualize it with a specific action are often limited. This method rotates the object to selectively visualize multiple data points as shadows.

BitCube is a new, simplified way of interacting with art and technology. Users can learn BitCube's programming logic without being PC-literate and easily make artistic works.

A novel thumb-sized force display that creates translational force sensations by using a sensory illusion that occurs when we perceive asymmetric vibration. The display produces haptic cues for helping pedestrian navigation and instructing hand movement by combining the force display with a position-sensing system.

This effective system for physical-based deformation transfer performs spring stretchiness ratio transfer on the models of abstraction levels, where the mass spring configurations are more consistent. The fine-level target model is reconstructed from a coarse-level model, thus completing deformation transfer.

A new approach to teaching middle-school children computational thinking concepts such as sequences, parallelism, or loops by creating choreography for a virtual character. Based on embodied pedagogy, movement and programming activities are combined, and students perform the motions together with the three-dimensional characters they program.

A method for enhancing the accuracy of color correction in stereo images by extracting relevant correspondence-pair data for color transfers using a modified stereo-matching algorithm and by compensating the color using weighted-sum value based on color differences of locally similar corresponding pairs.

A cone-tracing-based method to compute glossy and specular reflections in screen space.

This research concentrates on providing an innovative real-time, interactive gaming system called Cyberball to be played interactively in functional magnetic resonance imaging (fMRI) for studying empathy and social exclusion. The system incorporates virtual characters and proposes neuroscientific correlates of believability and emotional engagement with synthetic characters.
Invited to the First Round of the Student Research Competition

An anatomically accurate Android face capable of 35 degrees of freedom that aims to overcome the most significant mechanical shortcomings of the current Android technology for exploring human cognition.

Using an ultra-high-resolution image of an old picture scroll of Osaka (by Shunboku Ooka, 1745), interactive content for a Google Earth fly-through and an Augmented Reality Guide were developed and exhibited in public for regional promotions of Osaka.

Draco is a sketch-based interface that allows artists and casual users alike to add a rich set of animation effects to their drawings, seemingly bringing illustrations to life. The tool is built upon kinetic textures, providing continuous animation effects while preserving the unique, timeless nature of still illustrations.

This system summarizes racquet sports video automatically with an efficient viewing system that uses fast-forwarding.
Invited to the First Round of the Student Research Competition

A novel method that automatically finds various candidate poses suitable to the target environment and lets a user select the preferred pose.

This program extracts the topology of a hand-drawn cartoon to create a frame for animation. It improves a vectorization method developed in 2013 by researchers from Disney Studios, which was very accurate but too slow.

This realistic facial fattening and slimming simulation method uses a single frontal face image and is based on the skull structure obtained from MRI databases. The method reflects individuality based on the input’s skull structure and prevents the facial surface from penetrating into the skull.

In this gaze-based drawing-assistant system, patterns of eye movements during observation are projected onto the canvas to provide visual aids for what to draw, which allows different interpretations and drawing styles.

Obtaining global rendering effects such as shadows, reflection, and refraction during interactive rendering of normal maps using simple image-processing operations such as line-integral convolutions.

HaptoMIRAGE is an autostereoscopic display for seamless interaction with mixed-reality environments. The system can project 3D images in mid-air with a 180-degree wide-angle view, and up to three users can observe the same image from different points of view.

IllustStyleMap is a method for visualizing illustrations based on similarity of drawing style of authors. It first trains classifiers for author identification and then visualizes illustrations in 2D space.

Analysis of several implementations of a glasses-free tabletop 3D display that uses a hollow conical screen and tiny projectors. The image-display principle, characterized by using no obstructing apparatus on the table, employs optical sensors to recognize the region around 3D images.

A method for segmenting manga images interactively.
Invited to the First Round of the Student Research Competition

Understanding how different image modalities (line drawings, photographs) represent the same object is a key image-processing challenge. This poster presents a deep neural network that abstracts visual structure sufficiently well to produce congruent category maps for drawn, photographic, and rendered images, extracting “core” similarities that next-generation algorithms can exploit.

A novel method of rendering translucent materials in real time by acquiring translucent properties from real objects and then, by analyzing actual measurement, realizing real-time rendering of translucent materials.
Invited to the First Round of the Student Research Competition

The proposed method accomplishes 3D object categorization using metric-tensor and Christoffel symbols as a novel set of features and a polynomial-kernel-based SVM framework.

MOSS-xels is a novel slow-changing substantial display using blocks of Racomitrium Canescens. The moss pixels provide a light and dark contrast in appearance slowly and dynamically since, when water is absorbed, the leaves instantaneously open, and as the moss dries the leaves slowly close.

A technological framework that expands the hearing experience in the urban environment by marking its basic spatial elements and transforming them to sounds. The refined music notes are decoded back into urban elements. The resulting urban environment is a more balanced eco-system, viable and sustainable.

Capturing high-quality 3D models of insects is challenging. They are usually too small for laser or depth-camera-based systems, and techniques such as Micro CT scanning do not record color. This prototype system generates unprecedentedly high-quality natural-color 3D models of 3mm-30mm insects.

Simulation of deep ocean waves using a surface-only technique that inherently captures the inifinite depth of the ocean and handles solid boundaries by utilizing a novel boundary integral equation formulation and methodology.

This method of rendering and controlling stylized highlights uses light and view position to effect desired changes to the shape and position of a highlight . Another important feature is pseudo-normals, which simplify distracting highlights caused by original surface-normal details.

Pathfinder Vision assists future prediction to support short-term path planning by presenting images that consider physical interference between models of the surrounding environment and the current state of an operation target. The method alerts users to collisions before they occur.

With no prior assumptions about a camera's response, this new photometric calibration algorithm, AutoLum, automatically and directly measures the quantization boundaries of a camera with precision beyond the camera's own quantization levels and reduces photometric stereo angle error by 10x.

This project generates graphics using levitated small objects. It expands the acoustic-manipulation method by changing the distribution of the acoustic-potential field. The approach makes available many expressions, such as expression by materials and nondigital appearance, and these are levitated in mid-air.
Invited to the First Round of the Student Research Competition

This project's goal is to make it possible to create quasi-3D face rotation animation of 0~90° from frontal and profile images by controid interpolation and shape interpolation of a facial part.
Invited to the First Round of the Student Research Competition

Realistic behavior of spiders walking on a web is currently realized by animation loops. This proposed algorithm dynamically creates a natural walking behavior and can be performed on mobile devices in real time and without pre-rendered elements. The algorithm was finally integrated and evaluated in a mobile game.
Invited to the First Round of the Student Research Competition

This poster presents realistic rerendering of landscape photographs by extracting a view-dependent depth map from single-input landscape images for novel viewpoint renderings. The approach also relights landscape images by estimating the irradiance due to the sky in the input photograph.

Seeing around corners, in the dark, and through smoke is so far impossible with a mobile phone. This project demonstrates that with a microphone and speaker pair, such as the ones found in mobile phones, it is possible to sense the specular reflection of silent objects around occluding objects like walls.
Invited to the First Round of the Student Research Competition

An approach for determining stroke thickness in illustrations based on two-dimensional light position.
Invited to the First Round of the Student Research Competition

An interactive spatial augmented reality system that uses a single depth sensor for real-time object tracking.

Spheree is a personal, spherical, multi-projector rear-projected display with head-coupled, perspective-corrected 3D content for rich 3D interaction.

A perspective- and parallax-preserved image completion method for stereoscopic architectural images.

A proposed approach to converting existing anaglyph videos into stereoscopic 3D videos with full-color views. The approach builds on a temporal-coherence framework that also produces temporally coherent disparity maps and optical flow fields, which can be used for various post-processing tasks.
Invited to the First Round of the Student Research Competition

This poster describes light transport in transient-state and demonstrates new sampling strategies to adapt rendering algorithms to include the extra involved temporal dimension. The approach renders time-resolved phenomena, leading to animations that visualize light in motion and enable a better understanding of light transport itself.

Ubiquitous Substitutional Reality immerses users in an experience of both live reality and past reality through visual and audio feedback. Users are seamlessly transitioned to past reality when they interact with sensor-integrated furniture in a living environment.
Invited to the First Round of the Student Research Competition

This virtual fitting system uses 3D cameras (for example, Kinect for Windows) to construct customized body models for online shoppers and integrates offline physically based simulation and machine learning to simulate how garments deform on the virtual bodies in real time.
Invited to the First Round of the Student Research Competition

The Visual Simulation Laboratory application contains a visual-programming framework, with which users and programmers can easily construct flexible computational pipelines at application runtime on heterogeneous hardware platforms.

The current status of development of wearable displays that play an important role in the expanded 3D display platform. The optic system supports more natural 3D stereo effects that allow users to interact directly within near-body space.