Inter-frame synthesis between limited views

This study was started to improve functionality of devices which have imaging sensor. Methods have been applied to have more information about images taken from limited view angles by imaging devices. It's aimed to synthesis inter frames from two images taken from two different point of views. Square markers have been used to be able to create a mathematical relation between two different images. Automatic detection of the markers in images has let us to calculate positions of cameras. Key points have been detected on base images to synthesis of inter frames. Some different key point detection methods such as AKAZE, SURF, BRISK, ORB have been evaluated and compared. Detected points have been matched by Brute Force and FLANN based matching algorithms to calculate relations between two different base images. Inspection and evaluation of success of these two algorithms can be found in related sections of this document. Depth information has been calculated from two base images is needed to generate inter frames. Epipolar geometry rules have been used for this calculation. 3D interpolation methods based on well-known points have been used to get estimation of image areas that have no information to approach meshes of scene objects. Therefore, 3D structures of scene objects in images have been reconstructed as much as possible from base images that have been taken from different point of views.