Optical feature extraction
Divya Sharma, Shashikant Patil, Prarthna Patel, Utsavi Pathak · 2010
In recent times Optical signal processing have gained attention to produce feature spaces in parallel and the parallel architectures are attractive tools for optical pattern recognition. Most of Optical Feature extractors are very much different from Optical Signal Processing Systems which are specific to each of Application. Optical Feature Extractors are different and flexible in various applications. They are applying different discriminant functions to the output vectors produced by the system are useful in digital post processing. Here we are addressing Optical Feature Extraction. In Optical Signal Processing the corelators and Pattern recognition various filter function are required. This is not the case for OFE which computes certain geometrical properties of an input object. By feature extraction we can determine Object's location, scale and orientation. They can provide ability to achieve distortion invariance because they have potential for a larger number of classes. We are here proposing introduction to feature extraction and associated Linear Algebra for Processing. Here we are discussing methods to achieve distortion invariance and to reduce the dimensionality of the feature space. Also we are highlighting optical feature space study and concept of high dimensionality space. The use of Parallel Optical Feature Extractors was discussed in conjunction with Feature Space.