Intrinsic Image Transfer for Illumination Manipulation

Abstract

This paper presents a novel intrinsic image transfer (IIT) algorithm for illumination manipulation, which creates a local image translation between two illumination surfaces. This model is built on an optimization-based framework consisting of three photo-realistic losses defined on the sub-layers factorized by an intrinsic image decomposition. We illustrate that all losses can be reduced without the necessity of taking an intrinsic image decomposition under the well-known spatial-varying illumination illumination-invariant reflectance prior knowledge. Moreover, with a series of relaxations, all of them can be directly defined on images, giving a closed-form solution for image illumination manipulation. This new paradigm differs from the prevailing Retinex-based algorithms, as it provides an implicit way to deal with the per-pixel image illumination. We finally demonstrate its versatility and benefits to the illumination-related tasks such as illumination compensation, image enhancement, and high dynamic range (HDR) image compression, and show the high-quality results on natural image datasets.

Semi-Sparsity for Smoothing Filters

“…Recently, no-reference approaches based on statistical models have also shown promising success in predicting the quality of images. As suggested in [54], we use Tone Mapped Image Quality Index (TMQI) [55], Integrated Local Natural Image Quality Evaluator (IL-NIQE) [56] and Neural Image Assessment (NIMA) [57] for evaluation. The TMQI [55] index is a full-reference assessment method between HDR image and the output LDR image, in which Structural Fidelity (SF) and Statistical Naturalness (SN) are considered to provide an objective quality assessment.…”