In this paper, we address the problem of no-reference quality assessment for digital pictures corrupted with blur. We start with the generation of a large real image database containing pictures taken by human users in a variety of situations, and the conduction of subjective tests to generate the ground truth associated to those images. Based upon this ground truth, we select a number of high quality pictures and artificially degrade them with different intensities of simulated blur (gaussian and linear motion), totalling 6000 simulated blur images. We extensively evaluate the performance of state-of-the-art strategies for no-reference blur quantification in different blurring scenarios, and propose a paradigm for blur evaluation in which an effective method is pursued by combining several metrics and low-level image features. We test this paradigm by designing a no-reference quality assessment algorithm for blurred images which combines different metrics in a classifier based upon a neural network structure. Experimental results show that this leads to an improved performance that better reflects the images' ground truth. Finally, based upon the real image database, we show that the proposed method also outperforms other algorithms and metrics in realistic blur scenarios.

No-Reference Blur Assessment of Digital Pictures Based on Multifeature Classifiers

A smart-focusing technique includes identifying an object of interest, such as a face, in a digital image. A focus-generic classifier chain is applied that is trained to match both focused and unfocused faces and/or data from a face tracking module is accepted. Multiple focus-specific classifier chains are applied, including a first chain trained to match substantially out of focus faces, and a second chain trained to match slightly out of focus faces. Focus position is rapidly adjusted using a MEMS component.

Rapid auto-focus using classifier chains, mems and/or multiple object focusing

In this paper, we describe a novel multifocal projection concept that applies conventional video projectors and camera feedback. Multiple projectors with differently adjusted focal planes, but overlapping image areas are used. They can be either differently positioned in the environment or can be integrated into a single projection unit. The defocus created on an arbitrary surface is estimated automatically for each projector pixel. If this is known, a final image with minimal defocus can be composed in real-time from individual pixel contributions of all projectors. Our technique is independent of the surfaces' geometry, color and texture, the environment light, as well as of the projectors' position, orientation, luminance, and chrominance.

Multifocal projection: a multiprojector technique for increasing focal depth

Blind blur assessment for vision-based applications

The type and extent of blur affect image quality and therefore its evaluation. This paper presents an accurate method for blur identification and parameter estimation from one image without a priori knowledge. The key idea of the proposed method is to perform Fourier transform of logarithm spectrum to detect the periodic blur pattern in cepstrum domain instead of spectral nulls. Accordingly, the estimation of blur parameters is more accurate and robust to noise. The experimental results validate the accuracy of the proposed method.

Blind Image Blur Identification in Cepstrum Domain

We propose a simultaneous out-of-focus blur estimation and restoration technique for a fully digital auto-focusing system. The proposed auto-focusing system can make in-focused images by using only digital image processing techniques, and it requires neither infrared light or ultrasound nor a focusing lens assembly operated by a motor. As a result, it can provide savings in the volume and cost of the imaging system, which contains the proposed auto-focusing system. The proposed system first divides the input image into multiple sub-images or blocks, determines the set of blocks which contain linear edges, estimates the step response across the edge for each block in the selected set, computes the one-dimensional step response by averaging the step responses, estimates the two-dimensional point spread function (PSF) from the averaged step response, and finally provides in-focused image by an image restoration filter based on the estimated PSF. There is a limit on the amount of out-of-focus, which can be recovered by the proposed auto-focusing system. The proposed system operates under the assumption that an input image contains at least one piece-wise linear boundary between an object and the background. In spite of the above mentioned limitations, the proposed system can make an acceptable quality of focused image by using only digital signal processing.

Simultaneous out-of-focus blur estimation and restoration for digital auto-focusing system

111. PROPOSED IMAGE DEGRADATION SYSTEM Digital image restoration based on estimated PSF Ab8tractIn order to increase accuracy in estimating the point spread function (PSF) of out-of-focus blur, we subdivided the input image and then averaged PSFs from each sub-image. The PSF of each subimage is estimated by differentiating step response along the appropriately estimated edge direction under assumption that an image usually has piece-wise linear boundary between an object and background. Once the averaged PSF is obtained we can use any kind of image restoration filter according to a speciflc application requirements, such as computational complexity or processing time.

http://ieeexplore.ieee.org/iel4/5570/14904/00678418.pdf

Simultaneous Out-of-focus Blur Estimation And Restoration For A Digital Auto-focusing System

With the advancement of autonomous driving technology, a driving status monitoring system gains increasing attraction in the mobility field. In this paper, we present a drowsy driving detection system using intelligent recognition of driving status. The proposed method first detects the driver's face, and then extracts face landmarks for face analysis. Drowsiness driving detection is performed by analyzing the following two characteristics based on the extracted landmarks: i) eye condition analysis based on pupil recognition and ii) head movement analysis based on the head pose. The proposed method used an efficient algorithm for real-time analysis, and the performance of the proposed method was evaluated using in-house datasets.

Drowsy Status Monitoring System based on Face Feature Analysis

Optimal path planning is a key component in recent autonomous driving technology. Many navigation companies have developed services for predicting and navigating the optimal path by receiving information from users in real-time. However, these global path-based methods cannot predict the local, refined path due the resolution limitation of the GPS. In this paper, we present an object detection and lane detection system to predict the optimal local path using point cloud data(PCD) acquired by an autonomous dataset collection vehicle.

Point Cloud-Based Lane Detection for Optimal Local Path Planning

This paper describes a view-window control technique for intelligent video communication systems in which a local user can arbitrarily control the remote view-window and can magnify the selected area in the transmitted image without image degradation.

Sang Rae Park

Papers

Simultaneous out-of-focus blur estimation and restoration for digital auto-focusing system

Simultaneous Out-of-focus Blur Estimation And Restoration For A Digital Auto-focusing System

Drowsy Status Monitoring System based on Face Feature Analysis

Point Cloud-Based Lane Detection for Optimal Local Path Planning

Fully digital view-window control for intelligent video communication systems