ABSTRACT
The recent development of technological applications has made it inevitable to replicate human eyesight talents artificially and the issues requiring particular attention in the ideas of solutions increase in proportion to the number of applications. Facial classification in admittance restriction and video inspection is typically amidst the open-ended applications, where suitable models have been offered to meet users' needs. While it is true that subsequent efforts have led to the proposal of powerful facial recognition models, limiting factors affecting the quality of the results are always considered. These include low-resolution images, partial occlusion of faces and defense against adversarial attacks. The aspect of the input image, the verification of the presence of face occlusion in the image, the motive derived from the image, and the ability to fend off adversarial attacks are all examined by the RoFace formal representation of the face, which is presented in this paper as a solution to these issues. To assess the impact of these components on the classification/recognition accuracy, experiments have been conducted.
ABSTRACT
A new linear-features detection method is proposed for extracting straight edges and lines in synthetic-aperture radar images. This method is based on the localized Radon transform, which produces geometrical integrals along straight lines. In the transformed domain, linear features have a specific signature: They appear as strongly contrasted structures, which are easier to extract with the conventional ratio edge detector. The proposed method is dedicated to applications such as geographical map updating for which prior information (approximate length and orientation of features) is available. Experimental results show the method's robustness with respect to poor radiometric contrast and hidden parts and its complementarity to conventional pixel-by-pixel approaches.
