ABSTRACT

Image detection is process of detecting the images of real world. It is highly researched area form past few decades because of its day by day increasing demand in different
fields such as, security, medical, tile industries & biometrics. One may notice that surveillance cameras especially in malls & public places to record the activities of peoples!itis an important application of image detection known as security which may be done via surveillance cameras. Another vital application is medical imaging, includes - CT-SCAN, X-ray, ECG & EMG especially. It is also used in tile industries where the reference design is already stored in system 's memory, which is compared with the images of tiles manufactured by the machine & with the help of image subtraction approach it produces the output image showing the deviation in manufactured tile design from the reference design.
Several image detection techniques have been proposed yet along with some advantages and limitations. This thesis presents the comparative analysis of 6 image detection techniques such as SURF. FAST, BRISK, HARRIS, MinEIGEN, & MSER using Matlab 20 l 5a. These image detection techniques are analyzed in different cases
& scenarios, to simplify the analysis three different Scenarios (Scenario I : Single Image. Scenario 2: 3 Card Image & Scenario 3: Cluttered Scene) are considered along with four cases (Simple Case, Rotated Case, Scaling Case, & Illumination Case). These 4 cases are used in scenario 1, 2 and in 3 to analyze the performance of techniques and to compute the Accuracy in terms of Detected Feature Points.