In the field of augmented reality (AR), geometric and photometric registration is routinely achieved in real time. However, real-time geometric registration often leads to misalignment (e.g., jitter and drift) due to the error from camera pose estimation. Due to limited resources on mobile devices, it is also difficult to implement state-of-the-art techniques for photometric registration on mobile AR systems. In order to solve these problems, we developed a mobile AR system in a significantly different way from conventional systems. In this system, captured omnidirectional images and virtual objects are registered geometrically and photometrically in an offline rendering process. The appropriate part of the prerendered omnidirectional AR image is shown to a user through a mobile device with online registration between the real world and the pre-captured image. In order to investigate the validity of our new framework for mobile AR, we conducted experiments using the prototype system on a real site in Todai-ji Temple, a famous world cultural heritage site in Japan.

/pdf/mobile-ar-using-pre-captured-omnidirectional-images-1zf3n8bqfm.pdf

Mobile AR using pre-captured omnidirectional images

ISMAR 2014 : IEEE and ACM International Symposium on Mixed and Augmented Reality , Sep 10-12, 2014 , Munich, Germany

/pdf/indirect-augmented-reality-considering-real-world-irkt9aedt4.pdf

Indirect augmented reality considering real-world illumination change

The use of digital videos in criminal investigation and civil litigation has become popular, this is due to the advancement of embedded cameras in handheld devices such as mobile phones, PDA’s and tablets. However, the content of digital videos can be extracted, enhanced and modified using inexpensive and user friendly video editing software, such as; Adobe Photoshop, Sefexa, etc. Thus, the influx of these video editing softwarelead to the creation of serious problems that are associated with the authenticity of digital videos by making their validity questionable. In order to address these problems, two approaches for the authentication of digital videos were proposed by digital forensic researchers. The approaches are either active or passive. Active approaches are the earliest form of video authentication techniques; an active approach is based on digital watermark technology that is used for video authentication and ownership verification. A digital watermark is a hidden digital marker embedded in a noise tolerant video signal. However, the problem with the active approach to video authentication is that it can only be applied in limited situations and it requires the use of a special hardware. Moreover, an authorized person responsible for the watermark insertion can tamper with the video before inserting the digital watermark. Furthermore, techniques for encryption can be used to prevent an unauthorized person from tampering with the content of the video, however, these encryption techniques donot prevent the file owner from tampering with his own video. This limits the ability of digital watermark to ensure authenticity in digital videos. In response to these limitations, passive approaches were introduced. Passive approaches rely on the behaviour of features embedded in a video for forgery detection purposes. Thus, the aim of this doctoral study as a contribution to the field of digital forensic is to develop techniques based on selected video features that can be used to detect tampering of a digital video. In this study, passive forensic techniques are proposed to detect (1) Digital video inpainting forgery, and (2) Chroma key forgery in digital videos. Each of these techniques focus on the specific features that can be used to detect that kind of forgery. Firstly, a technique for the detection of video inpainting forgery is proposed using the statistical correlation of hessian matrix features extracted from the suspected video. Secondly, another technique is proposed for the detection of chroma key forgery in a digital video using the statistical correlation of blurring features extracted from the suspected video. Results from these experiments conducted have proven that hessian matrix features can effectively be used to detect video inpainting forgery with 99.79% accuracy whilst the blurring feature can effectively detect chroma key forgery in digital videos with 99.12% accuracy.

Passive video forgery detection using frame correlation statistical features / Aminu Mustapha Bagiwa

This paper presents a novel inpainting method based on structure estimation. The method first estimates an initial image that captures the rough structure and colors in the missing region. This image is generated by probabilistically estimating the gradient within the missing region based on edge segments intersecting its boundary, and then by flooding the colors on the boundary into the missing region. The color flooding is formulated as an energy minimization problem, and is efficiently optimized by the conjugate gradient method. Finally, by locally replacing the missing region with local patches similar to both the adjacent patches and the initial image, the inpainted image is synthesized. The initial image not only serves as a guide to ensure the underlying structure is preserved, but also allows the patch selection process to be carried out in a greedy manner, which leads to substantial speedup. Experimental results show the proposed method is capable of preserving the underlying structure in the missing region, while achieving more than 5 times faster computational speed than the state-of-the-art inpainting method. Subjective evaluation of image quality also shows the proposed method outperforms the previous methods. key words: image inpainting, image completion, texture synthesis, image structure

Fast and Structure-Preserving Image Inpainting Based on Probabilistic Structure Estimation

Indirect augmented reality (IAR) employs a unique approach to achieve high-quality synthesis of the real world and the virtual world, unlike traditional augmented reality (AR), which superimposes virtual objects in real time. IAR uses pre-captured omnidirectional images and offline superimposition of virtual objects for achieving jitter- and drift-free geometric registration as well as high-quality photometric registration. However, one drawback of IAR is the inconsistency between the real world and the pre-captured image. In this paper, we present a new classification of IAR inconsistencies and analyze the effect of these inconsistencies on the IAR experience. Accordingly, we propose a novel IAR system that reflects real-world illumination changes by selecting an appropriate image from among multiple pre-captured images obtained under various illumination conditions. The results of experiments conducted at an actual historical site show that the consideration of real-world illumination changes improves the realism of the IAR experience.

Addressing temporal inconsistency in indirect augmented reality

Toward a really useful navigation system, utilizing spherical panoramic photos with maps like Google Street View is efficient. Users expect the system to be available in all areas they go. Conventional shooting methods obtain the shot position from GPS sensor. However, indoor areas are out of GPS range. Furthermore, most urban public indoor areas are crowded with pedestrians. Even if we blur the pedestrians in a photo, the photos with blurring are not useful for scenic information. Thus, we propose a method which simultaneously subtracts pedestrians based on background subtraction method and generates location metadata by manually input from maps. Using these methods, we achieved an underground panoramic view system which displays no pedestrians.

Pano UMECHIKA: A Crowded Underground City Panoramic View System

In our daily life, we commute between home and office. Even if we leave at around the same time and on the the same route, the arrival time might not be the same due to the difference in walking speed, traffic conditions and trains to ride. Therefore, it is useful to develop a movement assisting system which can give the user timely pace-making advice. In order to give such advice automatically, the system has to know an accurate progress of the movement, the method of transportation and its mobility characteristics; walking speed, routes and traffic congestion characteristics. Hence, we propose a movement assisting system that proactively checks user's movement progress and gives timely advice upon automatically created personalized plan. Utilizing user's daily life log data of movements, it extracts key features as the places the often visited at and methods of transportation between them, and estimates arrival time to the destination according to their movement characteristics. In the case of movement by bus, our system could reduce 25% of arrival time estimation error compared with the conventional methods.

Yusuke Satonaka

Papers

Pano UMECHIKA: A Crowded Underground City Panoramic View System

Lifelog-based active movement assistant system