Eş Zamanlı Sıkıştırıcı Algılama ve Kriptografi

---

Thesis Type: Doctorate

Institution Of The Thesis: Yıldız Teknik Üniversitesi, Graduate School of Natural and Applied Sciences, Turkey

Approval Date: 2016

Thesis Language: Turkish

Student: Ertan Atar

Principal Supervisor (For Co-Supervisor Theses): Lale ÖZYILMAZ

Open Archive Collection: AVESIS Open Access Collection

Abstract:

Compressive Sensing was first developed by Emmanuel Candes, Stanford University in 2004 when dealing with problems of magnetic resonance imaging. He put forward that test data could be obtained independently from Nyquist criteria. In daily life, audio, video, and radar signals used in many applications such as video streams are sparse or compressible in general. Signals are first measured without considering compressibility, and then converted to a sparse signal, for example, by using a small number of transform coefficients containing most of the information. On the other hand, compressive sensing (CS) deals with how to measure the sparse signal in a compressed domain to start with, and how to retrieve the orginal signal by using CS theory. According to the Compressive sensing theory, if data/image compressing methods are not efficient or resolution of the data is not too poor, original data could be rebuild. Aim of the method is increasing the number of the sensor to take signal and convert to analog data to digital data. Compressive sensing uses a (usually random) linear projection matrix to convert the sparse signal to a measurable signal which is very much compressed. In this way, new display systems, cameras, sensor networks, and many applications such as the design of analog-digital converter can be developed. Recovery of original sparse signal requires compatibilty with the fundamental CS measurement theory. Cryptography is the name given to methods used in encryption and decryption of information. Reliability, data integrity, studies on mathematical methods of dealing with information security, issues such as authentication are important research topics in cryptography. Overall cryptography is divided into three types: symmetric, asymmetric, and hybrid cryptography. In this study, a compressive sensing algorithm called Orthogonal Matching Pursuit (OMP) is used together with symmetric, asymmetric and hybrid optical cryptography. In the method proposed, simultaneous compressed detection, symmetric and asymmetric cryptographic hybrid encryption / decryption is accomplished. Thus, both compressive sensing and data size reduction (reducing the number of sensors) is combined with cryptography by using CS detection matrix together with twice complex, random and distributed phase+amplitude keys as used in optical transform encryption and decryption methods with security-enhanced cryptography features. Cryptoanalysis results generated were successful.