%0 PDF %T Human Visual System Based Object Detection and Recognition and Introduction of Logarithmic Local Binary Patterns for Face Recognition. %A Mandal, Debashree. %8 2017-04-19 %R http://localhost/files/qb98ms319 %X Abstract: This thesis aims at incorporating logarithmic image processing and the human visual response which is based on Weber's law into various image processing applications. Human Visual System (HVS) has been used in this thesis for image decomposition. Specifically HVS based image decomposition has been applied towards the development of a novel framework for object detection and recognition systems. Also Logarithmic Image Processing (LIP) has been used towards the development of novel feature vectors. Logarithmic Image Processing (LIP) replaces the linear arithmetic (addition, subtraction, and multiplication) with a non-linear one, which more accurately characterizes the nonlinearity of computer image arithmetic and is consistent with the Weber's Law and the saturation characteristics of the human visual system. Two systems have been developed. One of which detects eyes from facial images after performing morphological operations. It has been shown that extracting features from HVS decomposed images followed by a feature fusion results in a better rate of detection than when extracting features from the original image alone. This has also proved effective in images that have shadows near the eye region. This thesis also presents a novel approach to the problem of face recognition that combines the classical Local Binary Pattern (LBP) feature descriptors with image processing in the logarithmic domain and the human visual system. Particularly, we have introduced parameterized logarithmic image processing (PLIP) operators based LBP feature extractor. We have also used the human visual system based image decomposition to extract features from the decomposed images and combine those with the features extracted from the original images thereby enriching the feature vector set and obtaining improved rates of recognition. Experiments have clearly shown the superiority of the proposed scheme over classical LBP feature descriptors.; Thesis (M.S.)--Tufts University, 2012.; Submitted to the Dept. of Electrical Engineering.; Advisor: Karen Panetta.; Committee: Sos Agaian, and Suchin Aeron.; Keywords: Electrical engineering, and Engineering. %[ 2022-10-12 %9 Text %~ Tufts Digital Library %W Institution