Depth Image Super‐Resolution Reconstruction Based on Fractal Encoding
Qifeng Niu, Yibo Zhao · IET Image Processing · 2026
ABSTRACT In response to challenges such as edge blurring and pseudo‐artefacts in the super‐resolution reconstruction of depth images, this paper proposes a depth‐image super‐resolution reconstruction algorithm based on fractal encoding. The algorithm leverages the contraction mapping and collage theorem of fractals to explore structural similarities between domain and range blocks in depth images, establishing transformation relationships between similar blocks. Utilising the resolution‐independent characteristic of fractal encoding, the algorithm employs the fixed‐point theorem of fractals for super‐resolution reconstruction of any initial depth image. In addition, an edge‐guided interpolation algorithm is used to predict residual information, which is then fused with the high‐resolution depth image to enhance the quality of the resulting high‐resolution depth image. Experimental results demonstrate that the algorithm improves both the peak signal‐to‐noise ratio and structural similarity for reconstructed super‐resolution depth images. Furthermore, the reconstructed depth images exhibit complete edge structures, clear overall contours, and richer details.