Spatial Light Modulation Pattern Generation by Golomb Code for Single Pixel Imaging

Silicon Focal Plane Array (FPA) image sensors with millions of pixels are used in modern digital cameras for array detection. Single-pixel imaging modality is beneficial where array detection is unavailable and fewer measurements are required to reconstruct the image. A spatial light modulator is the foremost element of this system. It is used in this system to add time-varying structured patterns before or after the object is photographed. In single-pixel imaging, both random and orthogonal patterns can be used as controlled patterns. To achieve high accuracy larger number of measurements of the random binary patterns is required, while orthogonal patterns like Hadamard and Origami patterns provide high reconstruction rates. Therefore, the reconstruction of images using deterministic patterns is quite useful. As Golomb coding is used as a lossless coding technique, the objective of this work is to generate spatial light modulation patterns through Golomb coding. To reconstruct an image using the Golomb coding pattern, an input image is masked with the pattern to obtain measurements. Through these measurements, the image is reconstructed. The performance evaluation of the proposed modulation patterns is compared with the Hadamard and Origami patterns for the reconstruction of different input images. Though the reconstructed images are with low resolution, this pattern offers a lot of potential for improving image quality for reconstruction.