17 February 2014 | Harbinder Singh, Vinay Kumar, Sunil Bhooshan
This paper presents a novel technique for exposure fusion using Weighted Least Squares (WLS) optimization to refine weight maps. The method aims to combine details from images captured at different exposure settings while avoiding visual artifacts. The proposed approach uses computationally simple texture features and color saturation measures to generate weight maps, which are then used to control the contribution of each input image in the fusion process. Unlike traditional HDR reconstruction and tone mapping, the method directly generates a well-exposed fused image suitable for conventional display devices. The technique is particularly well-suited for multifocus image fusion, offering advantages such as ease of implementation, high-quality compositing, and detail enhancement without introducing artifacts. The effectiveness of the proposed method is demonstrated through simulation results compared with various existing single-resolution and multiresolution techniques, showing superior performance in preserving details and maintaining local contrast.This paper presents a novel technique for exposure fusion using Weighted Least Squares (WLS) optimization to refine weight maps. The method aims to combine details from images captured at different exposure settings while avoiding visual artifacts. The proposed approach uses computationally simple texture features and color saturation measures to generate weight maps, which are then used to control the contribution of each input image in the fusion process. Unlike traditional HDR reconstruction and tone mapping, the method directly generates a well-exposed fused image suitable for conventional display devices. The technique is particularly well-suited for multifocus image fusion, offering advantages such as ease of implementation, high-quality compositing, and detail enhancement without introducing artifacts. The effectiveness of the proposed method is demonstrated through simulation results compared with various existing single-resolution and multiresolution techniques, showing superior performance in preserving details and maintaining local contrast.