Evaluating the robustness of image matting algorithm

Abstract

In this study, the authors propose a method to calculate the consistency of alpha masking to assess the robustness of the matting algorithm. This study evaluates consistent alpha masks based on the Gaussian–Hermite moment in combination with gradient amplitude and gradient direction. The gradient direction describes the appearance and shape of local objects in the image, and the gradient amplitude accurately reflects the contrast and texture changes of small details in the image. They selected Gaussian blur, pretzel noise, and combined noise to destroy the image, and then evaluated the consistency of the original alpha mask and noise alpha mask. To determine the robustness of the matting algorithm, they assessed the degree of consistency of the alpha mask using three different evaluation levels. The experimental results show that noise has a greater impact on the performance of the matting algorithm, which shows a decreasing trend as the noise level in the image deepens. In noisy images, the traditional matting algorithm exhibits better robustness compared to the recently proposed trap matting algorithm. Different matting algorithms present different adaptations to different noises.

Graph convolutional network‐based image matting algorithm for computer vision applications

“…Thus, these algorithms including the recent deep learning-based ones [10,11] are inevitably sensitive to the subtle overlapping region between the foreground and background components with various defect-related drawbacks. In the work of [27], Yuan, Li, and Fan proposed to compute the consistency of alpha masking to evaluate the robustness of the matting algorithms. To create an accurate trimap, Yuan and Li [29] presented an ant colony algorithm to yield the boundary information of the foreground objects.…”