In this paper, a contrast enhancement method called Multi-Histogram Mapping and Fusion (MHMF) is proposed for color images. Histogram analysis based method has been successfully applied to contrast enhancement in some applications, but they are hard to enhance the dark and bright regions simultaneously for back-light images. To solve this problem, the color barycenter model (CBM) is extended to separate the color image into lightness and chroma components. Then multi-histogram mapping (MHM) is used to map the lightness component of one single color image into several new lightness components with different contrast. These new components are divided into several patches and the best patches are selected by calculating image entropy. Finally, the selected patches are fused to create the enhanced image, and mix-Gaussian filter is applied to remove the sharp transition. The experimental results show the effectiveness of proposed method comparing with other state-of-the-art methods.

View full abstract

This paper presents an efficient constant-time algorithm for Gaussian filtering and also Gaussian derivative filtering that provides a high approximate accuracy in a low computational complexity regardless of its filter window size. The proposed algorithm consists of two key techniques: second-order shift properties of the Discrete Cosine/Sine Transforms type-5 and dual-domain error minimization for finding optimal parameters. The former enables us to perform filtering in fewer number of arithmetic operations as compared than some state-of-the-art algorithms without integral images. The latter enables us to find the optimal filter size that provides the most accurate filter kernel approximation. Experiments show that the proposed algorithm clearly outperforms state-of-the-art ones in computational complexity, approximate accuracy, and accuracy stability.

View full abstract

We present about a characteristic and the performance of the rotated character recognition using eigen-subspace for Japanese characters of 3,133 categories. At first, a few advantages of recognition experiments by noise-free character images are described. Next, it is shown that the multiple-projection method is more effective than the simple projection in low dimension, also the angle is precisely estimated within ±2 degrees. And the tendency of misclassification is considered for both projection methods. Next, although we use 36 character images as learning data, their sufficiency is shown experimentally. Furthermore, it is shown that very high recognition performance is obtained even when 3,133 Japanese categories of popular Mincho and Gothic fonts are used. Finally, when camera images are used, the deterioration of the recognition performance is improved. Also, it is shown that the inclination angle of a document can be estimated by this method without using layout analysis.

View full abstract

We propose a practical method that efficiently acquires light transports from arbitrary objects by using push broom-type projector illumination; vertical-strip and horizontal-strip light patterns are used. The combinations of vertical-strip and horizontal-strip camera responses allow our method to obtain light transports between each overlapped projector pixel and the camera response detected in common. Our method significantly reduces the number of camera response images needed for acquiring all light transports compared to brute-force scanning. Tests of the proposed method using real images show that it obtains more accurate light transports from various objects than brute-force scanning, and the resulting light transport matrix faithfully reproduces the original images. We demonstrate a relighting test based on the light transport matrix and confirm that our method yields relighted images that closely approximate the actual images captured by a camera.

View full abstract

Although FDM 3D printers have been getting popular, it is still difficult for novice users to determine basic parameters such as deposition directions and model subdivision. This paper describes an optimization method that determines the optimal model subdivision and deposition direction. The cost function is based on the supporting structure volume and the number of regions to adhere, which can be computed from multi-layered depth maps on GPU. The optimization can be efficiently achieved by the dynamic programing. Experiments showed that the method can improve object surface smoothness and reduce working time of support removal.

View full abstract

In this paper, we propose an algorithm to realize an immersive free-viewpoint experience, whereby audiences can see real sport scenes from any viewpoint, from a single moving camera. Homography matrix estimation and object extraction are two of the most important processes for synthesizing free-viewpoint video, since a world coordinate position of each object needs to be calculated in every frame based on object textures and homography matrix between captured frame and 2-dimensional world coordinate model of sports coat. Most conventional methods only use static cameras whose homography matrices and background models do not change. However, broadcasters usually utilize moving pan-tilt-zoom cameras in sporting events, and therefore, those methods cannot be directly applied to the events. In order to extend the range of application of free-viewpoint video technology to be used in sport live broadcasting, our proposed method estimates homography matrices semi-automatically by identifying reliable corresponding feature points between video frames, and also extracts the precise textures of objects using estimated homograph matrices. Experimental results revealed that the proposed method successfully estimated the precise homography matrices compared to the conventional method. Moreover, it was also confirmed that the proposed scheme contributes to further improvement in the experience of free-viewpoint video since the textures of multiple objects were successfully extracted.

View full abstract

An accurate (reliable) motion estimation (ME) algorithm using block matching for sequences including spatially periodic patterns is proposed and evaluated in comparison to conventional algorithms. Mismatch between accurate motion vectors (MVs) and MVs obtained by block matching (BM) which minimizes BM distortion is evaluated. BM distortion caused by MV sampling error (DVSE), which causes ME errors, is discovered. A new ME algorithm which uses the maximum value of DVSE (maxDVSE) is proposed. As for the proposed algorithm, maxDVSE is calculated for each block as a threshold of accurate vector candidates, and all candidate vectors for each block in a frame are stored. Accurate vectors are selected from the candidate vectors by finding so-called "anchor vectors" (namely, vectors at a unique point on spatially periodic patterns). The MVs are selected in a process starting from one block and propagated to neighboring blocks. Simulations are performed and the results show that average precision of the proposed ME algorithm is 98.9%, which is superior to that of other conventional algorithms.

View full abstract

A reproduction method for the 11 frontal channels of a 22.2 multichannel sound system (22.2 ch) over a direct-view display is investigated. NHK is planning to broadcast 8K Super Hi-Vision with 22.2 ch in a pilot broadcasting in 2016. A loudspeaker array frame has been developed for integration into the direct-view display as a home reproduction system and is used as a reproduction device of 22.2 ch. Localization of the frontal channels reproduced by the real loudspeakers is subjectively evaluated in terms of localization accuracy by estimating the difference from the intended localization for an 85-inch display. Three reproduction methods to synthesize the channels on the display are proposed and evaluated by comparison with a real loudspeaker placed at an exact channel position. Appropriate reproduction methods for the 11 frontal channels of 22.2 ch were decided from the results of a subjective evaluation.

View full abstract

Interference phenomena caused by multilayer films generates beautiful scenery. When rendering the phenomena in computer graphics, we should take into account both spectral distributions and global illumination. The paper proposes a method for rendering optical phenomena caused by multilayer films based on wave optics under global illumination environment. The method is an extension of a multilayer-film model, where the traditional model is based on a local illumination model and has limitations of rendering a scene with caustics and inter-reflection of light. In the proposed method, the multilayer-film model is reconstructed taking into account global light transport to render optical phenomena caused by interference of light under global illumination environment.

View full abstract

This paper presents a method to detect and track a paper map and overlay virtual data over it. First, we take a picture of the map and identify the road colors using our road area extraction method. Using the color information we can then extract in real time the roads from a video and use the position of the intersections to identify the map from a database without any further user input. The main issue is to get reliable intersection positions in real time even though the map can contain a lot of overlapping data. Finally, using the location of the intersections we can track and augment the map with virtual information. Our method enables the real time matching, 3D tracking and augmentation of a paper map based on a database that can contain several thousands of intersections coordinates.

View full abstract

In this paper, we propose a simple and effective technique, named extended joint bilateral filter, for reducing color bleeding which is defined as a smearing of the color between areas of strongly contrasting chroma in lossy-compressed images. The color bleeding comes from not only quantization but also sub-sampling in the YCbCr 4:2:0 color format. In order to deal with the two problems simultaneously, we extend the joint bilateral filter formulation by referring all of Y, Cb and Cr components based on a correlation between luma and chroma. As a result, our method achieves both noise reduction and sharpness enhancement for chroma without color edge blurring. Experimental results show that our method is effective in terms of both objective quality and subjective quality.

View full abstract