A Blur Classification Approach Using Deep Convolution Neural Network

A Blur Classification Approach Using Deep Convolution Neural Network

Shamik Tiwari (School of Computer Science, University of Petroleum and Energy Studies, Dehradun, India)
Copyright: © 2020 |Pages: 19
DOI: 10.4018/IJISMD.2020010106

Abstract

Computer vision-based gesture identification is designed to recognize human actions with the help of images. During the process of gesture image acquisition, images suffer various degradations. The method of recovering these degraded images is called restoration. In the case of blind restoration of such a degraded image where blur information is unavailable, it is essential to determine the exact blur type. This article presents a convolution neural network model for blur classification which categories a blur found in a hand gesture image into one of the four blur categories: motion, defocus, Gaussian, and box blur. The simulation results demonstrate the improved preciseness of the CNN model when compared to the MLP model.
Article Preview
Top

1. Introduction

Gesture identification is an area in computer vision with the aim of inferring human gestures via image analysis. Gestures can devise from movement of some part of body like face or hand images (Mitra & Acharya, 2007). Hand Gesture image are commonly used method for most of Human Computer Interaction (HCI) applications (Rautaray & Agrawal, 2015). An efficient human computer interface system required to recognize actions in real time. However, image-based hand gesture recognition is challenging issue owing to image degradation factors such as blur, noise, poor illumination etc. Large range of broadly available hand-held electronic devices like mobile phones, which offers an inbuilt camera are usually available these days. Enabling these devices with the competence to recognize gestures is a cheap and portable solution than the system with a conventional static camera (Nicholas, Marti, van der Merwe, & Kassebaum, 2017). A gesture image captured by camera phone can be transmitted to the system for decoding purpose. With such a mechanism, end-user will benefit from communication delivered by the mobile phone with the complete mobility. Integrating mobile phone with these systems is a revolutionary step for HCI environment.

The availability of camera phones offers a mobile platform for gesture recognition relatively than the use of the conventional camera, which suffers with mobility. Unfortunately, sometimes the deprived quality of the images captured by digital cameras makes it difficult to truly recognize gestures. Identifying gesture from images captured by common imaging devices is complex task because of restrictions of the integrated imaging system and the processing competencies of the device. These camera phones generally have lens of poor quality than the dedicated imaging cameras (Eisaku, Hiroshi, & Lim, 2004; Thielemann, Schumann-Olsen, Schulerud, & Kirkhus, 2004).

Image blurring is often a major issue, which degrades the any gesture recognition system. It is very challenging to recognize right gesture from a degraded image as shown in Figure 1. Image blurring is very unavoidable in an imaging system due to limitations of lens quality and other factors (Joseph & Pavlidis, 1994; Selim, 2004). Though good quality camera phones are reachable which provides high quality image with good resolution, the large number of users with low quality camera phone are still available.

Most of the times motion and defocus blur degrades the performance of mobile camera based HCI systems. Image restoration is highly desirable since blur diminishes the sharp features of image, which are highly desirable for any image analysis task including gesture recognition. Image restoration approaches are categorized into non-blind and blind restoration. Blind restoration of images where no prior information about blur is available, it is necessary to guess blur type before restoration.

Figure 1.

(a) A sharp image having simple white background, (b) degraded version of image (a), (c) A sharp image with complex background and (d) degraded version of image (c)

IJISMD.2020010106.f01

The works is discussed in seven sections including the present one. Section two offers literature review and Section three discusses four blur models that are considered in this work. Section 4 and 5 gives blur classification strategy and overview of machine learning models respectively. Section 6 deals with simulation and results. Lastly, conclusion is given in Section 7.

Top

2. Literature Review

In case of blind restoration of there is no uniform method is available which is applicable to all types of blurred images to identify the blur parameters. It is essential to know the type of blur in advance before application of any Point Spread Function (PSF) estimation method. Most of the previous research works emphases on categorization of image as sharp or blurred image. Very few researchers have focused on blur classification separately, which is more desirable in case of blind restoration.

Complete Article List

Search this Journal:
Reset
Open Access Articles: Forthcoming
Volume 12: 4 Issues (2021): Forthcoming, Available for Pre-Order
Volume 11: 4 Issues (2020): 3 Released, 1 Forthcoming
Volume 10: 4 Issues (2019)
Volume 9: 4 Issues (2018)
Volume 8: 4 Issues (2017)
Volume 7: 4 Issues (2016)
Volume 6: 4 Issues (2015)
Volume 5: 4 Issues (2014)
Volume 4: 4 Issues (2013)
Volume 3: 4 Issues (2012)
Volume 2: 4 Issues (2011)
Volume 1: 4 Issues (2010)
View Complete Journal Contents Listing