Object Detection in Smartphone Using Android

ISSN: 2394­3122 (Online) Volume 2, Issue 1, January 2015 SK International Journal of Multidisciplinary Research Hub Research Article / Survey Paper / Case Study Published By: SK Publisher Object Detection in Smartphone Using Android Mahesh Waghmode1 Bhagyashree K2 Department of Computer Engineering, JSPM College of Engineering Hadapsar, Pune, India Department of Computer Engineering, JSPM College of Engineering Hadapsar, Pune, India Bhagyashri D3 Chaitali S4 Department of Computer Engineering, JSPM College of Engineering Hadapsar, Pune, India Department of Computer Engineering, JSPM College of Engineering Hadapsar, Pune, India Abstract: Object detection is the technology used in computer to detect the object means save the information related to that object in the database. In the technology object is extract & defect object in an image which is implemented on the android platform. For implementing this technology image processing algorithm is used. In the further study of the object detection technique the (SIFT) scale invariant feature transform (SURF) is used. This is speed up Robust feature this are good method which yield high quality features. The smart-phones for user and that’s why it is used to implement this application. Keywords: Object detection, Smartphone, Android. I. INTRODUCTION Our goal is to build an android application for object recognition in mobile phone application for visually impaired users. Android is a development platform for the mobile application which is having the maximum market share among the mobile technologies. Detecting objects in a given image is the first step in image tagging applications. The information to be associated with the image is tagged to the detected objects. Thus user can get the information as and when required. So far object detection is done by various methods for desktop applications. Identification of objects in an image on android platform is not fully developed. The present study is focused on detecting objects in an image which later can be used for various other purposes. Objects once detected can be saved and is available for further use. In the present study image processing algorithms are considered for detecting objects. The blind and the visually impaired face diverse kinds of life challenges that normally sighted people take for granted. As far as out-door activities are concerned the blind indicate difficulties in safe and independent mobility depriving them of normal professional and social life. Then the issues dealing with communication and access to information are pointed out. Here a significant help is offered by software applications for computers and touch-screen devices equipped with speech synthesizers that enable browsing the internet and access to text documents. This application is developed only for android users as the name suggests. Now a day android is the fastest growing operating system for the smart phones and hence this application is profoundly developed for android users on. 1.1Propose system In this paper the object detection is proposed in following stages: Object is captured in mobile and detected after that related information of object is store. In scene detection the whole scene is detected and the information of that scene is tagged with the detected image of scene. Motion detection is process in which moving object is detected. © 2015, SK Publisher All Rights Reserved 14 | P a g e Mahesh et al., SK International Journal of Multidisciplinary Research Hub Volume 2, Issue 1, January 2015 pg. 14-19 1.2 Scope In this object detection system Image processing algorithm is used. » Blur algorithm » Gray scale algorithm » RGB to HSV conversion algorithm » Edge detection algorithm » Thresholding algorithm » Blob detection algorithm II. RELATED WORK Following algorithms are used in the system. 2.1 Blur Algorithm There are two types of blur Color blur and gray scale blur .Blurring means removing the noise in the image for this windowing is used. In Gray scale blurring the whole image is scan by width and height such as 3*3 or 5*5 for each pixel and noise in image is remove and image become more clear. In color blur first step is RGB separation and converting the image into gray scale image after this blurring is done Fig 2.1.1: Normal Image Fig 2.1.2: Blur Image © 2015, SK Publisher All Rights Reserved ISSN: 2394‐3122 (Online) 15 | P a g e Mahesh et al., SK International Journal of Multidisciplinary Research Hub Volume 2, Issue 1, January 2015 pg. 14-19 2.2 Gray Scale Conversion Algorithms In first step of gray scale conversion RGB separation is necessary RGB is three basic colors Red Green and Blue are separate out after that one variable is used to store the average of colors and image is converted into gray scale. 2.3 RGB to HSV Conversion HSV is detail information of the image .H is Hue means basic color of the image S is saturation means the concentration of the image V is value means brightness of the image. First step of the RGB to HSV conversion is to separate out the basic color then find out the min and max value after that in one temporary variable store the Max-Min value and compare that value with HSV such as H equal to 0 equal to 0and s equal to 0 and set pixels if the HSV is equal to 0. 2.4 Edge Detection Algorithm In Edge detection algorithm the edges of the main object is detected so that from whole scene the main object is recognize. For this Sobel edge detection algorithm is used in this algorithm two matrix are used Gx Gy for they are gradient of x and y. This algorithm dose not receive color image it only receive gray scale image. Formula for calculating gradient 2.5 Thresholding algorithm Thresholding is the simplest method of image segmentation. binary images is created through Thresholding i.e. image with only black or white. It is usually used for feature extraction where required features of image are converted to white and everything else to black. Fig 2.5.1:Simple Image © 2015, SK Publisher All Rights Reserved Fig 2.5.2: GrayScale Imag ISSN: 2394‐3122 (Online) 16 | P a g e Mahesh et al., SK International Journal of Multidisciplinary Research Hub Volume 2, Issue 1, January 2015 pg. 14-19 Fig 2.5.3: Threshold Image 2.6 Blob detection In the area of computer vision, blob detection refers to visual modules that are aimed at detecting points and/or regions in the image that differ in properties like brightness or color compared to the surrounding. There are two main classes of blob detectors (I) differential methods based on derivative expressions and (ii) methods based on local extreme in the intensity landscape. With the more recent terminology used in the field, these operators can also be referred to as interest point operators, or alternatively interest region operators (see also interest point detection and corner detection). 2.6.1 Steps in Blob detection 1. If a region has no higher neighbor, then it is a local maximum and will be the seed of a blob. 2. Else, if it has at least one higher neighbor, which is background, then it cannot be part of any blob and must be background. 3. Else, if it has more than one higher neighbor and if those higher neighbors are parts of different blobs, then it cannot be a part of any blob, and must be background. 4. Else, it has one or more higher neighbors, which are all parts of the same blob. Then, it must also be a part of that blob. III. STEPS IN IMAGE PROCESSING 3.1 Steps in object detection are » In object detection, the object is detected from the given image. » Firstly, the 32 bit original pixel is taken for reorganization. » The object from the given image is being recognized &analyzed. » The given object is then goes under blurring process. Here, the removal of noise in image is done. Image is converted to gray scale image. » After blurring, the grayscale Image. That is, it is converted to 8 bit separately. » Then the edges are made more visible I. edge detection. In edge detection, grayscale image is taken as input. » After thresholding is performed. The image is divided into foreground & background. The foreground is being highlighted and made more visible. » That is black=0 and white=1. Binary image is made from original image. » Blob detection is being performed. Where the blob (i.e. Collection of pixels) in the image where detected is assigned as object. © 2015, SK Publisher All Rights Reserved ISSN: 2394‐3122 (Online) 17 | P a g e Mahesh et al., » SK International Journal of Multidisciplinary Research Hub Volume 2, Issue 1, January 2015 pg. 14-19 Boundary detection is applied on object. Where, boundary of object is made visible for cropping. » After that, cropping is being performed. Here, we differentiate foreground from background & convert to HSV. » Till this process, object localization is done. » From the object being taken till the HSV, the image is taken &located on the android phone. » After, HSV being taken as input, histogram of the image of the image is being analyzed& stored. In histogram, the number of pixel for each color, feature extraction & represented in graph. » Then normalization is being taken i.e. range &color intensity is decreased. » Then object registration is made. Here, object is saved in serialization. » After registration, the object is recognized. » This are the basic & important step related to object detection. Fig: 3.2.1Flow Diagram of Object detection Fig: 3.2.2 Flow Diagram of Motion detection IV. CONCLUSION The image processing and recognition algorithms dedicated for blind users are proposed. Namely the color detector, the light direction detector and the object recognition algorithm. The developed software tools were implemented and tested on the smart phones equipped with a digital camera: HTC Desire HD, HTC Explorer and Sony Xperia S. We noted that performance © 2015, SK Publisher All Rights Reserved ISSN: 2394‐3122 (Online) 18 | P a g e Mahesh et al., SK International Journal of Multidisciplinary Research Hub Volume 2, Issue 1, January 2015 pg. 14-19 of these algorithms can depend on the quality of the built-in camera and image acquisition lighting conditions. The application is currently under tests among a number of blind users. References 1. Object Detection using FAST Corner Detector based on Smartphone Platforms. [Publish in 2011]. 2. Object recognition in a mobile phone application for visually impaired users. [Publish in 2013]. 3. An abandoned object detection system based on dual background segmentation. [Publish in 2014]. 4. An approach for Object Detection in Android Device. [Publish in 2014]. 5. H. Bay, T. Tuytelaars, L. V. Gool, “SURF: Speeded Up Robust Features,” in Proceedings of the European Conference on Computer Vision, 2006. 6. E. Rosten and T. Drummond, “Machine learning for high-speed corner Detection,” European Conference on Computer Vision, vol.1, pp. 430- 443. 2006 7. Hersh M., Johnson M. (Eds.) (2008) Assistive technology for Visually impaired and blind people, Springer, London. 8. Gill J. (2008), Assistive devices for people with visual impairments. In: Helal S., Mokhtari M. and Abdulrazak B. The engineering Handbook of smart technology for aging, disability, and Independence, J Wiley and Sons, Inc, Hoboken, New Jersey, pp. 163-190. 9. T. B. Moeslund, A. Hilton, and V. Kruger, “A survey of advances in vision-based human motion capture and analysis,” Computer Vision and Image Understanding, vol. 104, pp. 90-126, 2006. 10. L. Li and M. K. H. Leung, “Fusion of two different motion cues for Intelligent video surveillance,” Electrical and Electronic Technology, TENCON. vol. 1, pp. 19-22 Aug. 2001. 11. W. E. L. Grimson, C. Stauffer, R. Romano, and L. Lee, “Using adaptive © 2015, SK Publisher All Rights Reserved ISSN: 2394‐3122 (Online) 19 | P a g e