This paper presents a novel approach for real-time and proactive navigation in crowded environments such as event spaces and urban areas where many people are moving to their destinations simultaneously. The proposed approach is based on people flow predictions. As for the method, we newly introduce a spatio-temporal statistical method and experimentally compare it with the existing methods based on a matrix factorization approach using real data sets. Then, the effectiveness of the proposed navigation approach is verified by computer simulation using artificial people-flow data.

View full abstract

The content management methods based on the priority for the number of access, time and so on have been studied in cached video content distribution systems. However, since the variation related to the content access as a practical model has not been sufficiently considered so far, problems are still left for the variation of the number of access, and for the variation of the multiple-quality for each time zone. This paper proposes a cached video content management method by using priority estimation based on a content property. Considering the updating status of the cached contents, two kinds of contents, i.e., a regular content and a periodical updating content, are defined, and each content priority is used. Its priority is estimated by using the number of access in the reference intervals and the ratio between two qualities. The 1day model is used as an evaluation model and the variation models of the number of access and the quality are defined. An LFU method is used as a conventional method (CM), and the simulation experiments to compare the cache efficiency are performed, between CM and the proposed method based on these models. Consequently, since the hit ratio of cached contents in the proposed method is found to be 12–22 percent greater than the ones in CM, the proposed method is judged to be more superior than CM.

View full abstract

A projecting technique of 3-D holographic recovered images with the use of the mist screen employed thin bubble film has been studied. This technique has the fluctuation-reducing property of the reconstructed images, which has been an important problem to be solved in the mist-screen process, however, a detailed study of the permissible durability of the thin film has been strongly required since the displaying period of the projected images in the screen is limited. For this reason, in this paper, we studied a constructing technique of the screen and the improving effectivity of this screen in the projecting process of 3-D holographic recovered images. As a result, we have seen that the displaying period of the projected images in the screen is improvable in this process more than in the previous technical processes.

View full abstract

Along with the progress of computer generated hologram, and the realization of the interference fringes of light with easiness, application of three-dimensional images, such as a hologram, into growing cell phone is expected. Moreover, since the holographic transmission requires a high-speed optical wireless communication, application of IrSimple infrared communication protocol having been utilized for a simple terminal is considered as promising. In this paper the creation of computer generated hologram, and the communication system applying the IrSimple protocol to visible light communication system are studied, and the experimental results are described.

View full abstract

To reduce the loads of a video content distribution system (VCDS), a video distribution method using caching is used. In particular, in order to improve the efficiency of VCDS, video content management (VCM) methods, which use the priority of the number of access and the usage times, etc., had been studied. Moreover, since the data amount of independent video quality is large in VCDS for providing multi-quality, VCM methods using hierarchical data had been studied. However, There are some problems in such cases where the numbers of access are changed for elapsed time, and/or popular contents are frequently registered to be used. Thus, the author focuses on the environment of updating popular contents, and proposes a VCM method based on the numbers of access under the hierarchical quality-transition rule. In the proposed method, the quality of the registered content as new cached contents is assumed to be limited to the low quality, and the controlled capacity is used as cache capacity for new popular contents to improve the cache efficiency. By the experiments under the practical model of distributed contents, the variation models of the number of access and the variation models of the quality, the author verifies that the efficiency of the proposed method are maximum 6.67 percent greater than that of the conventional scheme in the access hit ratios for cached contents.

View full abstract

Recently, an image feature description calculating a co-occurrence between some pixels which are allocated at local region has been attracted an attention as the method which shows effectiveness in an object detection. However, there are some problems. A calculation cost and the number of feature dimensions of those methods tend to increase exponentially with respect to a feature description method focusing on single pixel. This paper proposed a multiple resolution CoHOG (MRCoHOG) feature description method which is able to suppress these exponential increases into a linear increase without incurring reduction in the classification accuracy. MRCoHOG can reduce the number of a feature dimension by calculating the co-occurrence only between the adjacent pixels, and maintain the classification accuracy by extracting a feature from plural low resolution images. In our classification experiments using the vehicle data set cropped from surveillance images of parking area and “INRIA person data set”, these results showed that the performance of MRCoHOG is equivalent to CoHOG's one.

View full abstract

Technical demands for extraction of significant components from spatial models are increasing as 3D sensor and its application technology had been developed and popularized. Conventional feature extraction for 2D image cannot be applied to depth images analysis, and is not suitable for 3D construction analysis because it extracts gradation between pixel values. In this paper, we proposed a novel method for extraction of characteristic elements from 3D point cloud according to normal vector distribution. In our numerical experiments, characteristics on each point were extracted by the proposed method with robustness to sensing noises, structural holes and irregular density.

View full abstract

The demand for character recognition in scene images has been growing. However, complex background, non-uniform illumination conditions and shadows are main obstacles for accurate detection and recognition of characters in scene images. For accurate detection, some methods that utilize a supplementary information given by a user with a touch operation using a device such as smart phones or tablets have been proposed. Based on this idea, in this paper, we propose a method for detecting character regions in scene image using auxiliary information given by a user. In the proposed method, the user will designate rough positions of characters in the image by an auxiliary lines and character regions are extracted based on the information. The effectiveness of the proposed method is experimentally evaluated with a public database. It achieves better result for character region extraction if the characters are with complex background and uneven brightness.

View full abstract

We have previously proposed the artificial fiber (AF) patterns in order to be able to hide information in printed documents. AF pattern uses the features of the medium (e.g., paper). It has some special characteristics of low visibility of the hidden information, print-scan robustness, device independent property and geometrical/rotational invariance. On the other hand it also has a weakness regarding foreground robustness and extraction threshold instability when using a camera as an input device. In the past several challenges we have proved foreground robustness only in case of adding AF pattern on foreground objects. It has been difficult to extract information in case of adding a foreground object on AF pattern by a distortion of its low frequency part. Also we found extraction threshold instability was caused by an uneven illuminance when using a camera. In this paper we proposed using the calculated intensity ratio of adjacent AF patterns for improving an uneven illuminance at camera usage then covering foreground objects with the AF pattern by detecting a contour of foreground objects for improving foreground robustness. We proved these proposals are very effective for improving weaknesses of the AF pattern by several experiments.

View full abstract

In this paper, we propose a shape reconstruction method for multiple human bodies with occlusion by human body parts matching and temporal pose data of human body parts. We define occlusions as three types: First type is a mutual-occlusion occurred between multiple human. Second type is a self-occlusion occurred on a single human body. Third type is a mutual-self-occlusion occurred in a case when the self-occlusion occurred in the mutual-occlusion. Firstly, the proposed method reconstructs human shape in the mutual-occlusion and the self-occlusion scenes by matching human body parts obtained in advance to human surface data out of these occlusion regions. Secondly, in order to reconstruct shape of human bodies in the mutual-self-occlusion, the proposed method estimates poses of human parts in the mutual-self-occlusion region by using temporal position-pose data of these parts and reconstructs shape of human bodies by applying these parts with estimated poses. In this paper, we quantitatively and qualitatively evaluated effectiveness of the proposed method through simulation and experiments with actual human bodies.

View full abstract

In the presentation of 3D volumetric holographic images, spatial projection technique with the use of mist screen seems to be very effective. However, in the previous study for this technique, complicated formed nozzles and a complicated optical system have been required. In this paper, we shall introduce a new mist screen display which enables us to present spatially volumetric holographic images in a more easier way. It has adopted only one fan which produces a stable spiral flow of the mist with the aid of mist-flow controller. Mist-flow controller is used for arrangement of the flow of the mist. It suggests that it gives us in a more easier way the volumetric holographic images in the mist screen in the spatially projected process, and it is also useful to suppress the fluctuation of the images generated when wind blows in the cross direction.

View full abstract

This paper deals with a problem where a robot identifies an object that a human asks it to bring by voice when there is a set of objects that the human and the robot can see. In this case, a human uses an expression which consistes of one or some attributes, such as color and name etc.. In this paper, we propose the method for the identication using color and object names. The multimodal information of speech and images are used for the identification.

View full abstract

In order to obtain information effectively from radar imagery, we propose a target extraction method based on the shape of reflected echoes and their motion. Our method exploits knowledge that targets are isolated regions and are stable temporally and spatially while conventional methods are based on intensity of reflected echoes. Morphological operations such as Quoit filter and segmentation enable the method to extract targets from images that include clutter in bad weather, weak echoes from small vessels and the regions divided by bridges and breakwaters. We conducted an experiment using datasets generated from radar imagery, which contain clutter and artificial structures. The result showed that our method outperformed a conventional one in the evaluation scale of recall, precision and F measure.

View full abstract