Takao Jinno

Multiple Exposure Fusion for High Dynamic Range Image Acquisition

A multiple exposure fusion to enhance the dynamic range of an image is proposed. The construction of high dynamic range images (HDRIs) is performed by combining multiple images taken with different exposures and estimating the irradiance value for each pixel. This is a common process for HDRI acquisition. During this process, displacements of the images caused by object movements often yield motion blur and ghosting artifacts. To address the problem, this paper presents an efficient and accurate multiple exposure fusion technique for the HDRI acquisition. Our method simultaneously estimates displacements and occlusion and saturation regions by using maximum a posteriori estimation and constructs motion-blur-free HDRIs. We also propose a new weighting scheme for the multiple image fusion. We demonstrate that our HDRI acquisition algorithm is accurate, even for images with large motion.

New local tone mapping and two-layer coding for HDR images

This paper proposes a two-layer High Dynamic Range (HDR) coding scheme using a new tone mapping. Our tone mapping method transforms an HDR image onto a Low Dynamic Range (LDR) image by using a base map that is a smoothed version of the HDR luminance. In our scheme, the HDR image can be reconstructed from the tone mapped LDR image. Our method makes use of this property to realize a two-layer HDR coding by encoding both of the tone mapped LDR image and the base map. This paper validates its effectiveness of our approach through some experiments.

Motion blur free HDR image acquisition using multiple exposures

The high dynamic range image (HDRI) acquisition method based on Markov random field model is proposed. By combining multiple exposure images shot with different shutter speed, we estimate the irradiance value for each pixel. Our method estimates displacements, occlusion and saturated regions, and by using them construct the motion blur free HDRI with higher quality than other existing methods.

Acquisition and Encoding of High Dynamic Range Images using Inverse Tone Mapping

A two layer coding algorithm for high dynamic range images is discussed. In the first layer, a low dynamic range image is encoded by a conventional codec, and then the residual information that represents the difference between an original and the decoded images in the first layer is encoded in the second layer, which realizes compatibility with conventional image file formats. Our method utilizes the approximation of an inverse tone mapping function that reduces the high dynamic range to a displayable range. Our algorithm significantly improves a compression performance, compared to conventional methods.

HDR image compression using optimized tone mapping model

In this paper, we propose a coding algorithm for High Dynamic Range Images (HDRI). Our encoder applies a tone mapping model based on scaled μ-Lawencoding, followed by a conventional Low Dynamic Range Image (LDRI) encoder. The tone mapping model is designed to minimize the difference between the tone mappedHDRI and its LDR version. By virtue of the nature of the model, not only the quality of the HDRI but also the one of LDRI are improved, compared with a state of the art in conventional HDRI compression. Furthermore the error caused by our tone mapping model encoding is theoretically analyzed.

HDR video tone mapping based on gamma blending

In this paper we propose a new tone mapping method for the HDR video. Two types of gamma tone mapping are blended to preserve local contrasts in the entire range of luminance. Our method achieves a high quality tone mapping especially for HDR video that has a nonlinear response to scene radiance. We examine the validity of our method through simulation and comparison with conventional work

Local covariance filtering for color images

In this paper, we introduce a novel edge-aware filter that manipulates the local covariances of a color image. A covariance matrix obtained at each pixel is decomposed by the singular value decomposition (SVD), then diagonal eigenvalues are filtered by characteristic control functions. Our filter form generalizes a wide class of edge-aware filters. Once the SVDs are calculated, users can control the filter characteristic graphically by modifying the curve of the characteristic control functions, just like tone curve manipulation while seeing a result in real-time. We also introduce an efficient iterative calculation of the pixel-wise SVD which is able to significantly reduce its execution time.

Detail preserving multiple bit-depth image representation and coding

Recent trends in high bit depth representation of images, including high dynamic range images, increases demand for multiple bit depth representation. When reducing the bit depth, detailed variations in the original image are lost through quantization. We propose a scheme for quantization using the bilateral based detail preservation. Our approach has an ability to precisely restore the original bit depth from the quantized image. We apply the multiple bit-depth representation for the image coding and show the validity of our method. We demonstrate the effectiveness of our approach through some experiences.

Advanced weighting scheme in the rational Remez algorithm for IIR digital filters

In this paper, we present a numerical method for the equiripple approximation of IIR digital filters. Many methods have been proposed for the design of such filters. One of the most practical methods is the rational Remez algorithm, in which the squared magnitude response of the IIR digital filter is approximated by solving an eigenvalue problem, and then the filter is obtained by factorizing it. Unlike the original Remez algorithm for FIR filters, it is difficult for the rational Remez algorithm to control the ratio of ripples between different bands. In the conventional lowpass filter design, for example, when different weights are given for its passband and stopband, one needs to iteratively design the filter by trial and error to achieve the ratio of the weights exactly. To address this problem, we modify the conventional algorithm and make it possible to directly control the ripple ratios. The method iteratively solves eigenvalue problems with controlling the ratio of ripples. Using this method, the equiripple solutions are obtained quickly.