Leong, Yong En (2021) Optical Encoding Based On Single Pixel Imaging. Final Year Project, UTAR.
Abstract
Single Pixel Imaging (SPI) is an imaging framework that utilises only one light detector instead of detector arrays required by conventional imaging sensors. This is achieved by sequentially applying a series of mask patterns and measures the light intensity using the single light detector. The measurements correspond to the mask patterns can then be used to reconstruct the image. SPI is inherently encrypted, since the light intensity collected cannot be related to the original image in any way or form, and can only be reconstructed based on the mask patterns used. Thus, the collected light intensities can be perceived as the ciphertext and the configuration of mask patterns is the key. Due to the inherent property of SPI, its encryption security and performance were studied. The image acquisition and reconstruction are essentially optical encoding and decoding process. Various SPI encoding and decoding schemes were investigated, which include Hadamard, random, Fourier, and Chaotic. The findings from the analysis of these existing schemes suggest the need of a new encryption method. The proposed method is essentially a double encryption that utilises the mixing of logistic chaotic maps using random index map at the first phase and Double Random Phase Encoding (DRPE) or Rivest–Shamir–Adleman Encryption (RSA) at the second phase. Assessment was performed from the perspectives of reconstruction quality and security analysis. Besides, the proposed method was also compared to the existing encryption schemes. The results show that the proposed method has average performance in term of image reconstruction quality at low sampling rate but has significant improvement in term of security. Therefore, the proposed method is able to overcome the shortcomings regarding the security of conventional SPI encoding schemes.
Actions (login required)
![[img]](http://eprints.utar.edu.my/style/images/fileicons/application_pdf.png)
