Data Compression for Single-Pixel Imaging Using Audio Codecs

Single-pixel imaging (SPI) reconstructs an object’s image via computational methods using a single-point photodetector and a sequence of mask patterns. A fundamental trade-off exists in SPI between the sampling rate and the fidelity of the reconstructed image: while increasing the number of patterns enhances image quality, it simultaneously scales the volume of the measured light intensity data. In this study, we propose a SPI data compression based on audio compression technology, taking advantage of the fact that light intensity data is inherently one-dimensional. Our results demonstrate that the Free Lossless Audio Codec algorithm achieves significantly higher compression ratios than conventional lossless methods based on LZ77 and Huffman coding, without compromising image quality. Furthermore, we show that lossy formats such as MP3 and Ogg Vorbis can achieve even higher compression ratios, albeit with minor degradation in the final reconstruction.