Image Shadow Removal Based on Multi-Scale Feature Fusion

Shadow is a common natural phenomenon. However, shadows in images not only affect the overall sensory quality of the image, but also affect the content recognition of the image. Many existing traditional and deep learning-based shadow removal methods process images directly in the RGB color space. However, since the brightness information and color information are mixed together in the RGB color space, it is not conducive to shadow removal. We convert the image to the LAB color space, separating the luminance and color information of the image, and use a multi-scale feature fusion module to extract information at different scales in the image, so as to better remove image shadows. Experiments demonstrate that our method outperforms several existing state-of-the-art methods on the ISTD dataset.