Shadows can originate from occlusions in both direct and indirectillumination. Although most current shadow removal research focuses on shadowscaused by direct illumination, shadows from indirect illumination are oftenjust as pervasive, particularly in indoor scenes. A significant challenge inremoving shadows from indirect illumination is obtaining shadow-free images totrain the shadow removal network. To overcome this challenge, we propose anovel rendering pipeline for generating shadowed and shadow-free images underdirect and indirect illumination, and create a comprehensive synthetic datasetthat contains over 30,000 image pairs, covering various object types andlighting conditions. We also propose an innovative shadow removal network thatexplicitly integrates semantic and geometric priors through concatenation andattention mechanisms. The experiments show that our method outperformsstate-of-the-art shadow removal techniques and can effectively generalize toindoor and outdoor scenes under various lighting conditions, enhancing theoverall effectiveness and applicability of shadow removal methods.