[in Japanese]

View full abstract

An extra line produced the reconstruction image by the simulation of conventional CGH by phase information. However, I realized the plasticity of the stable reconstruction image by taking a coefficient for only the amplitude in this article. I applied this result to communication and examined the data communication of many values using CGH with the information compression of the Frequency Division Multiplexing method.

View full abstract

A method of crime prevention wearing smartphones on bags such as school bags is proposed. Movies are taken by children using our smartphone devices and the child position is computed using GPS. The parents monitor their children for the crime prevention. Using our multimedia information, various defenses can be promptly implemented in emergency such as detection of suspicious individuals by face verification system and behavior analysis of persons appearing in the video.

View full abstract

A method of giving sticky note annotations to three-dimensional objects with image local feature is proposed. We take template images and query videos of a three-dimensional heart model by a camera and give annotations to the model. Using keypoints that we can get with image local feature, we estimate the positions of annotations in queries. We simulate that even if the model moves, annotations can track it.

View full abstract

A method for pseudo stereoscopic view of the painting using face recognition is proposed. Recognize face of painting by the face recognition, to calculate the distance and context of each persons, and to estimate the depth. Generate an anaglyph by moving the human area left and right with consideration for eyes parallax. Painting is not drawn from multiple perspectives can be simulated stereo. Results of our method using actual painting are shown in our experiments.

View full abstract

[in Japanese]

View full abstract

Currently, various mixed reality (MR: Mixed Reality) technologies are researched and developed, but existing MR systems are expensive, so it is not easy for ordinary users to obtain and operate them .Therefore, this study was performed MR system by using generic device was development. In addition, we make an investigation into objects on MR space using generic devices since there is no report discussing the user interface .This study , we developed an MR system by using a video see through head mounted display and real time synthesis of space images and virtual 3D objects by using AR markers. As a user interface for operating objects, GUI (Graphical User Interface) by controller and mouse operation, the interface allocating operation functions to buttons of controller and the interface by hand gestures using a depth camera was respectively implemented. In order to find the most useful interface, we performed an observer experiment on nine subjects, and found that the GUI by controller is the best interface.

View full abstract

With recent advances in Internet technology, consumers can easily get content uploaded on the network, edit it and create it as new content. However, at the same time, there is a demand for technology to protect the author's rights of the original content. As a technique to fill it, a method of controlling the secondary use of the contents according to the intention of the author by using the BLS signature has been proposed. In this method, in order to prevent secondary usage of content unintended by the author, authors can set signatures controlled on whether editing is permitted for each part of their contents. In this paper, we will implement the contents created by “MikuMikuDance” which can create and edit content, create a demo and display it.

View full abstract

The direction of handwriting movement is left on the handwriting as the direction of the stroke, the slope of linear portion of the stroke, and the change of direction in the curve of the stroke. Therefore, it is an important factor of writer identification. We aimed at the color circle with color-code over the 0 to 360 degrees angle range and studied a method of expressing the angle by the hue about the character composed of plural strokes. First, the character is decomposed to plural strokes. Then, thinning, calculation of the angle and coloring are executed to the strokes. Finally, the colored strokes are assembled into the colored character. Since the character, in and of itself, is colored, the observer can understands the difference in the characteristic of handwriting at a glance. This expression method is effective in visualization of the characteristic of handwriting in the attachment figure to the expert’s written opinion or report.

View full abstract

In recent years, the haptic modality seems to be important sensory system as much as vision in the field of virtual reality and design. However, the ways how visual and haptic information apply to texture recognition have yet to be fully elucidated. In this research, we conducted subjective evaluation experiments and used linear multiple-regressions analysts to model the relationship between visual, haptic and visuo-haptic impressions of synthetic resins texture. As a result, some of the models which include interaction terms (the visual rating multiplied by the haptic rating) showed high coefficients of determination.

View full abstract

In the field of computer graphics, it has been able to represent objects with CG images more realistically. However, there are few researches clarify visual impression given by objects and CG images by comparing them. In this study, we focus on emboss processing on surface of synthetic resin, and clarified visual impression given by real surfaces and generating texture images of synthetic resins. First, we measured surface characteristics of various emboss processing on synthetic resins, and generated texture images of them based on measurements. Second, we conducted subjective evaluation experiments in real surfaces and generating texture images. Finally, we clarified the relationship between the visual impression and their surface characteristics based on the results of evaluation. This suggests the possibility of feedback to improve the reality of CG.

View full abstract

We considered cached video content management methods in video content distribution systems with multiple quality. The linear approximation estimation method was proposed, and we considered a switching method using the estimation methods with time and property information (linear approximation estimation, secondary function approximation estimation, and exponentially approximation estimation) was proposed. By the simulation results, it is revealed that the proposed method is effective in cache efficiency to control the estimation errors for variation models.

View full abstract

This manuscript investigates the characteristics of the temporal filter that generates encoded video signal with high density temporal resolution.

View full abstract

I propose an invisible digital watermarking method for printed images. Copyright informations are embedded inside image data by using a pattern showing anisotropic green noise characteristics in the pixel space. In the frequency space, the spectrum of the original picture is confined inside the green noise spectrum. For this reason, separation is easy and extraction accuracy of watermark information is high even if the embedding strength (gain) is small. In the evaluation experiment, the watermark could be extracted from the printed image for gain≧0.125 . Watermark can be removed / re-embedded using a private key. This key is different each image, therefore provides high security system.

View full abstract

[in Japanese]

View full abstract

In this research, we propose a novel method of extracting 3D features using several geographic direction such as normal direction, minimum curvature direction (extrusion direction) and vertical direction. Based on machine learning (AdaBoost), it enables highly accurate classifier, suppressing erroneous discrimination on the object. As a specific task, we worked on tree trunk detection.

View full abstract

We propose an object proposal method that aims at precisely detecting all objects in depth images. Existing methods exploit RGB image boundary information to evaluate objectness of each candidate region based on the assumption that object boundary is a clue to evaluate the candidates. However, since RGB images inevitably include both object boundary and texture boundary, these methods cannot cover objects in images by small amount of candidate regions especially when many objects are shown in one image. We thus tackle this problem by designing a novel objectness measure from depth images which can easily extract object boundary as image boundary. Experimental evaluations demonstrate that our method can detect objects in test images by a smaller number of candidate regions than existing methods.

View full abstract

In recent years, due to the aging of the population, an increase in the number of care receiver and a shortage of care worker are becoming serious. Restricted to meal assistance, the number of care workers is short against the number of people requiring care, so there are care receiver who are waiting for meal assistance. In this paper, We propose a meal assistance robot that it is judged from RGB-D sequences whether food exists in the mouth of the care receiver, and when the mastication is over, the robot arm moves and safely carries the food to the mouth of the care receiver. The proposed method consists of judgement part whether food exists in the mouth, and a control part of a robot arm. In the judging part, state of meal is recorded using Kinect v2 which can acquire the coordinate of a face and a joint. Calculate the degree of opening of the mouth and posture angle from the obtained RGB-D sequences information, move the arm to the mouth of the care receiver when it is judged that the care receiver can start the meal. After that, it detects the presence or absence of mastication, judges that there is no food in the mouth when there is no movement, and carries the next food with the arm to the mouth. In the control part, the robot arm is rotated based on the coordinates of the mouth obtained by the judgment part, and the angle of each link is obtained by inverse kinematics. By sending each calculated angle to the microcomputer by serial communication, it controls so that the spoon attached to the tip of the robot arm is carried to the mouth. In the experiment, we repeatedly carry food to the mouth using the meal assistance robot by the above algorithm. After completing the experiment, we take a questionnaire to the subjects and evaluate the safety of robot from the summation results. As a result, the subject gave opinions indicating the effectiveness of this system.

View full abstract

In recent years, We are aware of many environmental problems, such as environmental destruction, reduction of wild animals. It is required to make use of autonomous robots for environmental monitoring to predict disasters and grasp situations. In this research, we consider image processing technology that contributes to environmental monitoring by WAMOT (Waseda Animal Monitoring Robot), an environmental monitoring robot developed for the purpose of creating environment and animal ecology maps. WAMOT uses a charging station to withstand long-term autonomous observation. When the battery of the robot runs short, it automatically returns to the charging station. As soon as charging is completed, repeat the operation of starting the observation again to grasp the surrounding environment. In this research, we aim to make this autonomous driving possible by using the visual information. We propose a method that enables stable self position estimation in irregular areas like forests by estimating the robot's state from the camera pose obtained by Large-Scale Direct Monocular SLAM (LSD-SLAM) and combining wheel odometry . As a result of the experiment, self - location estimation with higher reliability than wheel odometry alone could be performed.

View full abstract

In this paper, in order to perform automatic portrait generation, we consider a method of automatically identifying and extracting hair area from face image using deep learning. By learning a Convolutional Neural Network (CNN) using teacher data whose label image s of hair region s created manually are output s and front al facial RGB images are inputs , Semantic segmentation is performed to predict categories other than hair area or hair area. As a result, we obtained a prospect of automatically extracting the area of hair with high precision even for unknown face images.

View full abstract

We propose a method to estimate the traveling distance of the camera from the image that taken by a camera traveling on the central axis of the gas pipe. Define the trajectory by following the feature points on the inner wall of the pipe with the optical flow and estimate the distance between the start point and the end point of the trajectory using the pinhole camera model. The distance difference is taken as the traveling distance of the camera.

View full abstract

This paper presents several fundamental methods of estimating the time points of UAV crashes . For the purpose of protecting UAV during a crash event , we decide to use an airbag to provide UAV a soft cushioning and restraint Airbag is designed to inflate extremely rapidly then quickly deflate it is important to find out the timing of deploying the airbag. (Considering about using an automotive airbag, it is necessary to recognize the 0.1 second before crashes Therefore, our study is to compare the methods of measuring UAV’s flight height above the ground in each moment of the falling process by using camera and sensor.

View full abstract

Recently, people take selfies with their smartphones. For taking selfies, in-camera mounted inside the smartphone is often used. However, the in-camera has lower resolution than the out-camera mounted on the outside of the smartphone. One of the methods to improve resolution is super-resolution(SR). Among the SR technologies, nonlinear signal processing(NLSP) is a compact and light SR technology. However, the performance of NLSP on a smartphone has not yet been evaluated. We adjusted NLSP for selfies and evaluated the performance by subjective assessment. The assessment results proved that the image quality of the smartphone with NLSP is effective for selfies.

View full abstract

3D planes are figures with low geometric features, however, they have important applications in computer vision, such as object recognition, autonomous navigation and key points detection. Current methods are precise, but non-deterministic and slow; therefore, in the past we proposed a deterministic Hough Transform based method and confirmed faster speed and improved accuracy. Nonetheless, the number of parameters were big; thus, motivated on the sliding window methods in 2D, in this work we created a 3D Sliding Voxel approach to detect planar surfaces with less parameters, faster speed and higher accuracy.

View full abstract

Because of digital technologies, high-definition television (HDTV) has become a common video system. Commercial 4K TV broadcasting has started in Japan in 2015. 8K is also under developing, and the 8K TV broadcasting service is planned in 2020. 4K/8K are prospective media, and their high-resolution video has various possibilities for broadcasting, medical, publishing, and others. With regard to the video system equipment, cameras are very important because they determine the image quality, particularly resolution. However, a fine focus is more important than the capability of a camera. Professional HD/4K/8K cameras are not equipped with auto-focus systems because the focus areas should be controlled manually due to production-making requirements. The image is blurry if the focus is not fine. Because of operability, small and light-weight cameras are preferred. The size of the viewfinders is similar with that of the HDTV cameras. It is extremely difficult to adjust the focus with a small viewfinder even for a professional video camera person. Large bulky LCDs are used to adjust the focus even for outside broadcasting. They reduce the performance and obstruct mobility. Focus aid systems are proposed to cope with the focus issues. They are based on the linear signal processing method. However, these do not have tolerance against noise. Noise always appears if the lighting condition is not sufficient. In this paper, a nonlinear signal processing is proposed to fix the noise issue. The nonlinear focus assist method also shows a more accurate focus and smaller areas than the current systems.

View full abstract

Using polid screen of the tomography, I took the tomogram of the reconstructed image of the hologram and examined the evaluation method for the three-dimensional image.

View full abstract

We studied the holographic projection displaying system with the use of cavitation bubbles. As this result, highly bright projected images are observed together with the flicker noise suppressed smaller than before around them.

View full abstract

We report here an approach to the time-shard multiplex reconstructing technique which produces full colored projected holographic images. As this result, this technique seems to be useful for producing colored projection of holographic images.

View full abstract

This report intends to provide the latest information on International Standardization of image coding standards that are currently discussed in ISO/IEC JTC 1/SC 29/WG 1. This report also focuses on JPEG Privacy & Security activity. The activity aims to provide capabilities for secure image sharing with protection of privacy and security contained in image and metadata.

View full abstract

As of June 2017, A new international standard, called JPEG Pleno, is being discussed. This new standard is aimed to represent scenes that not only provides an image information about shooting point and direction but also about how it changes when observed from different positions and directions. This paper explains the establishment of the JPEG committee and the positioning of each organization in the international standardization organizations, and outlines the development process of the image standards at the WG 1 Committee. Next, introduce each image standard standardized by “JPEG” up to “JPEG Pleno,” and describe the features of these JPEG standards. After that, we discuss parameters necessary for plenoptic expression, introduce images by five kinds of representation methods showing plenoptic images, and outline procedure of image data processing using plenoptic processing model. Finally, we will introduce the establishment of JPEG Pleno standardization, the requirements of this specification, and the present of achievement of this standard development and future activities.

View full abstract

ISO/IEC JTC 1/SC 29 /WG 1 is standardizing still image coding technologies. The standards contain bi-level image coding methods and multi-level image coding methods. In this report, multi-level image coding standards such as JPEG, JPEG 2000, JPEG XR and JPEG XT are reviewed by their algorithm, easiness of implementation, and compression performance.

View full abstract

In the conventional image recognition system, image recognition and image coding are independent processing. Our proposal image recognition system transmits only a part of the hierarchical image from IoT devices (eye) to edge gateway (brain) and performs image recognition on edge gateway side. The power consumption of IoT devices are suppressed, the effective image recognition system for IoT are implemented. Only essential data are transmitted, image recognition are implemented with coarse image quality, the distributed image recognition processing is realized with low power consumption, finally in the cloud side, and is characterized by transmitting the object image of arbitrary shape. It’s possible to use many application.

View full abstract

This paper introduces still picture coding technology in video codec. Several key aspects of HEVC intra are highlighted. In addition, on-going study targeting higher compression efficiency toward future video coding are discussed.

View full abstract

The Organized Session named “To Where Still Image Coding will Head for?” , is composed of the introduction of recent still image coding standards, various comparison of JPEG family standards, and new trend of still image coding standardization. This panel session named “Future of Still Image Coding” is targeting to raise and discuss the related future subjects based on these presentations.

View full abstract

The latest trend of the Inkjet Continuous Feed Printing System and Post processing System to Outlook in “HID2017” at Switzerland in February

View full abstract

View of Digital Print from Canon I explain Canon’s activities with market information and case study.

View full abstract

Among all information transmission methods, it can be positioned as one of the optimum methods that customers desire.

View full abstract

Standardization activities are not merely activities to determine technical details. Corporate management should understand that standardization activities will affect the rise and fall of companies. Because system business has a difficulty to provide all the components by a single company, the necessity and importance of standardization activities are increasing further. Human resources for standardization activities should have not only technical knowledge but also political power. Corporates should develop such human resources in a long-term.

View full abstract

Introduction of background and initiatives situation of “ Three Action Plans for the Development of Human Resources Responsible for Standardization”, formulated on January 2017 by Working Group for Studying the Development of Human resources Responsible for Standardization under the Public-Private Strategic Council for Standardization.

View full abstract

A revision plan is proposed for “Skill standard －Evaluation for skills of human resource required for standardization” published in 2013. The plan reflects the new trends on standards/standardization as well as the new approach for skill buildings.

View full abstract

Survey and report of an ultimate comfortable space for human being through collaboration of music and image.

View full abstract

Fundamental characteristics of shell structures are going to be presented through simple experiments using chopsticks and eggs. Further, vibration damping of the structures and future research on this field are going to be presented.

View full abstract