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. Recently, JPEG XS low-latency lightweight image coding system and JPEG Pleno for Light Field image and Holography have been developed in SC 29/WG 1 to respond to market demand for VR/AR/MR applications.

View full abstract

In recent years, with the start of the new 4K8K satellite broadcasting and the spread of VR and immersive video, scenes requiring production and transmission of higher definition and higher data rate video are increasing. In such applications, a light-weight compression codec that can be compressed and decoded with low latency while maintaining the quality of high definition video as much as possible, and can be transmitted in a smaller bandwidth is required. In general, light-weight compression codecs are characterized by low latency of several lines, visually lossless, small circuit scale, and easy implementation on various platforms such as CPUs, GPUs, and FPGAs. This paper describes the standard of such light-weight compression codecs, and reports the conditions necessary to transmit full-specification 8K using light-weight compression codecs including JPEG XS, which was standardized in 2019.

View full abstract

Holography can record and reconstruct wavefront information from an object as two-dimensional image by interference with reference wave. Since the features of holographic images are quite different from ordinary two-dimensional images, conventional two-dimensional image coding cannot be applied efficiently. In this talk, I introduce the principles and features of holography, and some researches on coding.

View full abstract

It is easy to leave hidden risks because the traditional video surveillance system is unable to monitor all the operating scenarios due to equipment and environmental factors. This paper proposes to use the Linux platform to increase the intelligent video analysis server of the front-end substation, and introduce a large number of image identification and deep learning technologies, by combining the existing video surveillance system in the substation and affecting no existing system operation nor new investment. The paper studies the use of intelligent video analysis technology to establish a video security control and prevention mechanism for substation production areas, to integrate the existing monitoring and surveillance system of the substation, and to reserve API for data connection with automation systems such as the Two-sheets system(work guide sheet and operation sheet), safety production management system and information systems, so as to greatly improve the accuracy of system identification and add a stable alarm management platform. For this system, video detection can be conducted at the front-end by using the current intelligent video analysis server and the monitoring personnel can easily complete the work like alarm management and on-site evidence collection with the help of the alarm management platform.

View full abstract

Aim to “use information analysis technology to support a healthy longevity society for all generations in smart city”. What do you want to do, what can you do, what do you have to do? Even though it is impossible to realize all at once, we have to spend time thinking together. As a reference axis for that purpose, in this paper we tried to organize from the viewpoint of the prefix “Cyber”.

View full abstract

Smart city international standards provide common strategies for city management across the countries. In this paper, we introduce several ICT related international standards based on related works from the viewpoints of city indicators and architectures. We also explain our insights based on the experience as international expert.

View full abstract

Fundamentals on shell and spatial structures are explained and the damping ratios of the structures are discussed.

View full abstract

Smart houses in Japan have been developed in accordance with energy policy by the government, and this can be considered as one of IoT deployments. Recently many people are thinking about utilization of big data out of smart house. In this paper, latest policies about residential houses by the government are touched upon, then several research abstracts about smart houses that can make people happy are described.

View full abstract

Recently, the opportunity to combine and process several types of data in the image processing field has increased. In this research, we introduce machine learning by neural network as a method to associate information in different sensors and integrate multiple sensor data. We learn the relationship between the gray scale information and the depth information by machine learning, and thereafter, the depth information is restored from only the gray scale information based on the learning data.

View full abstract

[in Japanese]

View full abstract

Due to large amount of data, videos are encoded (data volume reduction) before transmission and storage. However, coded videos have degradations. One of them is called ‘block-noise’ that appears block shape degradation on the image. Block-noise can be usually reduced by a de-blocking filter that reduces the block boundaries. However, the conventional de-blocking filter has an issue that it cannot be applied to cropped video with the block-noise. Although the de-blocking-filter reduces the block noise, it also causes blur on the edges in the video. In this paper, we propose a novel de-blocking filter that can reduce block-noise for cropped videos with keeping the edges.

View full abstract

Due to advances of semiconductors, smartphones become common tools and security cameras are filled in downtowns. Smartphone pictures and recorded security camera images sometimes give us important information of crimes. Car license plate numbers are one of them because cars are often used in criminals. However, the recorded images/pictures are blurry and the license plate numbers cannot be readable. Although there is machine learning systems to improve resolution of blur numbers, their performances change depending on classifiers. In this paper, the performances of the classifiers are compared and the best one is reported.

View full abstract

The damage caused by the pirated films amounts to $1.3 billion every year. The pirated films are mainly created by re-shooting a screen in a theater or duplicating the official DVDs. Films are released in a theater. Before the official DVDs are released, 90% of the films are pirated in a theater and illegally made DVDs are sold. Although counter technologies are required, there is no effective proposals. The important requirement for the technology is that it does not degrade the image quality in a theater. It only degrades the re-shot videos. In this paper, we report a novel method to make it possible.

View full abstract

There have been various studies on image generation using the Generative Adversarial Network (GAN), and we have considered a game stage generation of 2D action games. In my previous study, we encoded to a simple image to decrease noises on the game stage, and the system was able to decrease noises on the generated game stage. In this study, we describe a method to restore encoded stages and generate playable game stages as a game. Specifically, we calculated the pixel values of the image generated by GAN an d the values defined by the conversion rules, converted them into objects according to the calculation results, and restored the playable game stage.

View full abstract

In sports training, efficient form acquisition is a common task for amateurs and professional players. In this paper, we aim to develop an interface that can acquire an ideal form in a short time using VR (Virtual Reality) technology. We choose squat training that is one of muscle strength training as a target and the Kinect v2 is used to reflect the joint coordinates of the subject in the human body model in the game development software Unity. In addition, five evaluation values of squat training is calculated from the joint coordinates of the human body model in Unity, and it is reflected on the color information of the spherical object. A bar that lights only when all items meet a certain evaluation criteria at the same time was also placed, and feedback is given as a measure of whether the subject moved on the ideal form. We observed how improved degree operation before and after the training and obtain the results shown prospects to the effectiveness of this system.

View full abstract

In this research, we focus on the joint work with people of automated forklifts in the manned warehouse and aim the efficiency of the warehouse work by selecting the traveling operation of the forklift for the operation of the worker on the aisle and avoiding the danger Let's say. The motion of the person to set was assumed to be two motions of walking motion and motion to stay on shelf work, and the purpose was to separate it from moving images. The operation was reproduced in the warehouse, and a motion image was acquired by Kinect. People are detected from moving images, skeleton information is acquired by HOG feature and Open Pose , and binary classification is performed by SVM. We verified the accuracy of motion identification and compared it with the HOG feature and OpenPose on the accuracy change due to the change of the frame amount of image data and verified its effectiveness.

View full abstract

Recently, Wu et. al have proposed a reversible contrast enhancement using digital watermarking. By reversible processing, it is possible to restore the original image from the contrast enhanced image. However, there is a problem that visible noise occurs in the contrast enhanced image This noise is caused by the pre processing to avoid the overflow or underflow of the luminance value. In this study, we propose a method to improve the image quality of contrast enhanced images by improving pre processing In our method, two adjacent bins with low occurrence frequencies are combined to make empty bins around the range boundary .

View full abstract

For the health management of honeybees reared for beekeeping, it is important to monitor the state of hexagonal cells , called brood cells, formed on nest frames In this study, we conducted a fundamental study on a method to automatically classify the condition of cells from image s of nest frames We defined state s of cells in six classes and applied a pixel based classification using CNN to several nest frames

View full abstract

When videos viewed by viewers have screen shake information, it is one of the important issue to prevent viewers from VIMS (visually induced motion sickness). A screen shake detection method using histogram of motion had been proposed to prevent viewers from screen shake, which was a factor of VIMS. However, there is a problem of the decrease of accuracy due to the relationship between motion information in original sequences and a screen shake. Then, in order to improve the accuracy, a screen shake detection method using motion histograms in the specified area is proposed, and the detection accuracy is evaluated by the simulation experiments.

View full abstract

We propose an invisible and reversible Green-Noise diffusion watermark using multi-level embedded pattern. The Green-Noise Diffusion represents robust resistance and high image quality even if embedded by strong “gain”. Using 4-level Green－Noise pattern, we can embed double volume of watermarks, furthermore introducing perturbation patterns, we can obtain higher security watermarking system from the attack by third party.

View full abstract

We examined the effect of overlapping with real objects on depth perception of virtual objects presented with optical see-through head-mounted displays (HMDs). In the experiment, virtual objects were presented with optical see-through HMDs, and there were two conditions of overlapping with real objects. Under the overlapping conditions, real objects were presented in front of the virtual objects. Under the non-overlapping conditions, the real objects were not presented. The depths of the virtual objects were measured by using the reference markers of real objects. The results obtained from the experiment revealed that individual differences were large in depth perception of virtual objects in mixed reality accomplished with optical see-through. These findings indicate the importance of developing the techniques adapted to such individual differences.

View full abstract

In recent years, many researches have been proposed to estimate the atmosphere of class and the degree of concentration of students in class from a video of the student s’ behavior. However, taking videos at specific student gathering places like classrooms is a violation of privacy. So, in this presentation, we try to analyze student s behavior from low resolution video which is hard to identify individually. The proposed method uses the Haar like classifier created from the smoothed image to detect the student's frontal face in the smoothed image. Then, from the change in the number of skin color pixels in the face area, it is determined when the student is facing the front. From experiment s , it was found that the analysis of the face direction of student s is sufficiently possible even from the smoothed image where the personal identification is difficult.

View full abstract

The high magnification zoom image of the 250 million pixel camera was confirmed to be the high-definition image that can clearly recognize the subject (passenger-plane, ship, car, traffic sign) within 20 km range. The developed video processing removes disturbance noise (atmospheric fluctuation, out of focus, haze) and changes the distant high magnification zoom image clearly. It was confirmed using image recognition function of AI that it was able to recognize a distant subject accurately than a usual surveillance camera.

View full abstract

8K is being developed for a broadcasting system. 8K is also applied to medical field as an endoscope. Using the 8K endoscope a laparoscope operation is popular among patients. However, a diameter 0.02 mm stitch is used. It is impossible to watch the 0.02 mm stitch with a 4K camera. Only 8K camera is the only tool to watch the thin stitch. However, fine focus is necessary to make it. In this paper, it is reported that non-linear super resolution technology make it possible to watch the stitch when the focus is not completely fine.

View full abstract

It is called “conformity assessment” to assess whether products or services are conformed to relevant standards. Conformity assessment is indispensable when any standards are utilized in business and society. The methods of conformity assessment include suppliers’ declarations of conformity, second-parties’ audits and certifications. Traditionally, many kinds of test charts had been developed in standards related to imaging technologies, and industry associations, academic societies and private companies had sold them for mainly being used for suppliers’ declarations of conformity and second-parties’ audits. But recently new certification systems have been introduced as imaging technologies have been used in many application fields including social safety. As this trend will continues in the future, it is important to consider a method of its conformity assessment simultaneously when a standard related to imaging technologies is developed.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract

Obun Printing Company has been undertaking research and development of braille printing and tactile printing using UV offset printing technology since 2011. Braille printing and tactile printing technologies using ultraviolet ray hardening resinous transparent ink are already commercialized by conventional printing technology. Observation of ease in tactile discrimination of braille and tactile printing by researchers has been reported, and JIS Standard (JIS T 9253: Performance and Test Method of Ultraviolet Ray Hardening Resinous Braille) has been established. Since braille printing and tactile printing using ultraviolet ray hardening resinous transparent ink can print braille and tactile graphic on printed matters for general sighted people without losing visibility of characters and patterns, they are expected to spread, as visually impaired people and sighted people can obtain common information. Braille printing and tactile printing using ultraviolet ray hardening resinous transparent ink by conventional printing technology are used for installed type facility maps; however, other than that, they seem not to be widely used for distribution-type printed matters. Braille printing and tactile printing by UV offset printing technology developed by the authors can quickly print, in bulk, printed matters (calendars, manuals, route maps, and facility maps) required for daily life of visually impaired people. The authors are aiming to spread these as universal design printed matters that allow visually impaired people to communicate with sighted people, making them an item to support the independence of visually impaired people. This paper reports the progress of our commitment to the research and development to date, and examines future promotion and enlightenment.

View full abstract

OrCam MyEye 2 is a small AI visual aid device that reads text information, recognizes people, things, barcodes, bills, colors, user gestures. You can hear the people's names and product names, bills, color, time, day of the week, and the date from the speaker at the ear. OrCam’s technology, is a great example of how AI can benefit humanity, with OrCam MyEye２, we can improve the lives of more than a billion people who are visually impaired, blind, or have a reading or learning disability such as dyslexia.

View full abstract

We have developed a Braille translation system, called BrailleMUSE, which provides Braille music scores from digital scores on WEB sites via the Internet. In this paper, the algorithm for translating Braille scores and the configuration of developed system are described. We have made its translation facility open to use around the world via an Internet since 2006. The translation availabilities of the system and the operation history for recent 15 years are also introduced.

View full abstract

Twenty-five years have passed since the development of music software B’Score started. This 　 software is used by both visually impaired and braille score translators. We will explain the development background, current situation, problems, and the future of Braille Music Notation.

View full abstract

In this paper, we introduce various fiber transport solutions for 5G and beyond network. We present the concept of optical self heterodyne system for generation and transmission of radio signals in the millimeter wave band. We experimentally confirm the satisfactory performance for broadband radio signal transmission in the 90 GHz band.

View full abstract

[in Japanese]

View full abstract

[in Japanese]

View full abstract