AutoFill: An Automatic Approach to Filling Panoramic Images
Su Wang, Mingwei Cao · 2024
Panoramic image filling stands as one of the foremost research topics in the field of image editing. The purpose of image stitching is to repair the blank area at the edge of the panoramic image, ensuring alignment with the original scene, and enhancing overall quality and usability. This paper proposes an efficient and high-precision automatic approach (called AutoFill) for filling panoramic images to defend the issue that missing edge information in panoramic images. The AutoFill method comprises the following key steps: first, calculate the binary image (Mask) for the input panoramic image; second, based on the binary image, calculate the area of the panoramic image that needs to be repaired; third, extract of pixel information from the neighborhood area surrounding the blank region to be repaired; and finally, use the extracted pixels to fill the blank area to produce the high-quality panoramic image. Multiple open-source data sets were used to evaluate the proposed AutoFill, and experimental results demonstrated its efficiency. Our method successfully completes the panoramic image filling task.