In the last decades, depth estimation from multi-view is treated as an ill-posed problem. This problem becomes severe with limited data such as sparse-view cases. However, with the presence of Convolutional Neural Network (CNN), recent learning-based depth estimation methods prove effectiveness on occluded and texture-less area where prior works still suffer on handling such issues. They utilize features from CNN layer for constructing cost volume and regress input volume with regression network. To overcome those concerns, we introduce a unique approach by combining hand-crafted and learning-based strategies. Specifically, we utilize the Normalized Cross-Correlation (NCC) cost volume, which is more robust to noise than simple L1 and L2 costs, to improve the photo-consistency between local patches. The entire construction pipeline is implemented by pyOpenCL to speed up the processing time. Finally, we employ the network that estimates depth by regressing handcrafted cost-based plane sweeping volume.
The recent works on the light field (LF) image enhancement are focused on specific tasks such as motion deblurring and super-resolution. State-of-the-art methods are limited with the specific case of 3-degree-of-freedom (3-DOF) camera motion (for motion deblurring) and straight-forward high-resolution neural network (for super-resolution (SR)). In this work, we proposed a framework that utilizes the deep neural net to solve LF spatial super- resolution and deblurring under 6-DOF camera motion. The neural network is designed with end-to-end fashion and trained in multiple stages to perform robust super-resolution and deblurring. Our neural network achieves superior results in terms of quantitative and qualitative performance compared to the recent state-of-the-art LF deblurring and SR algorithms.
A set of document image processing algorithms for improving the optical character recognition (OCR) capability of smartphone applications is presented. The scope of the problem covers the geometric and photometric distortion correction of document images. The proposed framework was developed to satisfy industrial requirements. It is implemented on an off-the-shelf smartphone with limited resources in terms of speed and memory. Geometric distortions, i.e., skew and perspective distortion, are corrected by sending horizontal and vertical vanishing points toward infinity in a downsampled image. Photometric distortion includes image degradation from moiré pattern noise and specular highlights. Moiré pattern noise is removed using low-pass filters with different sizes independently applied to the background and text region. The contrast of the text in a specular highlighted area is enhanced by locally enlarging the intensity difference between the background and text while the noise is suppressed. Intensive experiments indicate that the proposed methods show a consistent and robust performance on a smartphone with a runtime of less than 1 s.
This paper presents a Kinect–stereo camera fusion system that significantly improves the accuracy of depth map acquisition. The typical Kinect depth map suffers from missing depth values and errors, resulting from a single Kinect input. To ameliorate such problems, the proposed system couples a Kinect with a stereo RGB camera to provide an additional disparity map. Kinect depth map and the disparity map are efficiently fused in real time by exploiting a spatiotemporal Markov random field framework on a graphics processing unit. An efficient temporal data cost is proposed to maintain the temporal coherency between frames. We demonstrate the performance of the proposed system on challenging real-world examples. Experimental results confirm that the proposed system is robust and accurate in depth video acquisition.
KEYWORDS: High dynamic range imaging, Image acquisition, Point spread functions, Camera shutters, Image quality, Cameras, Optical engineering, Image analysis, Image sensors, Visual system
The authors propose a method to obtain a high dynamic range (HDR) image from multiple images with different exposure. Unlike conventional methods that use multiple images with different shutter speeds, the proposed method takes multiple images with identical and short shutter speeds but with different apertures. Consequently, the input low dynamic range image is less affected by scene change, while it has undesirable defocus blur due to different depth of field. In order to mitigate defocus blur of input images with larger apertures, we estimate the defocus map of each input image and use it as the spatially variant point spread function to deblur the image. Then, we extract the weight maps of input images, which are used to combine them to synthesize an HDR image. Our experimental results show that the proposed algorithm produces high-quality HDR images using a small number (typically three) of input images.
In this paper, we present a universal deblurring method for real images without prior knowledge of the blur source. The proposed method uses the transition region of the blurred image to estimate the point spread function (PSF). It determines the main edges of the blurred image with high edge measures based on the difference of Gaussians (DoG) operator. Those edge measures are used to predict the transition region of the sharp image. By using the transition region, we select the pixels of the blurred image to form a series of equations for calculating the PSF. In order to overcome noise disturbance, the optimal method based on the anisotropic adaptive regularization is used to estimate the PSF, in which the constraints of non-negative and spatial correlations are incorporated. Once the PSF is estimated, the blurred image is effectively recovered by employing nonblind restoration. Experimental results show that the proposed method performs effectively for real images with different blur sources.
We present a novel algorithm to remove motion blur from a single blurred image. To estimate the unknown motion blur kernel as accurately as possible, we propose an adaptive algorithm using anisotropic regularization. The proposed algorithm preserves the point spread function (PSF) path while keeping the properties of the motion PSF when solving for the blur kernel. Adaptive anisotropic regularization and refinement of the blur kernels are incorporated into an iterative process to improve the precision of the blur kernel. Maximum likelihood (ML) estimation deblurring based on edge-preserving regularization is derived to reduce artifacts while avoiding oversmoothing of the details. By using the estimated blur kernel and the proposed ML estimation deblurring, the motion blur can be removed effectively. The experimental results for real motion blurred images show that the proposed algorithm can removes motion blur effectively for a variety of real scenes.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.