An Extended Exposure Fusion and its Application to Single Image Contrast Enhancement

Simulated Exposure Fusion

“…This section presents SEF [8], a method generating a set of images with simulated exposure from a single input. This model simply applies different multiplicative factors to the input image.…”

“…Although the simulated images in [8] are obtained using a set of contrast factors, the authors call them under-and over-exposed. Indeed, using a simplistic model of the digital images formation (and notably omitting the non-linearities of the acquisition process), simulating images with a shorter or longer exposure amounts to multiply them by a constant.…”

“…Let us denote by u the input image, by x a pixel, and by u a simulated image. To artificially increase the exposure time, the authors of [8] use…”

“…using λ k < 1, is not desirable, because they would have lower contrast and still no information in the clipped pixels. The authors of [8] therefore suggest to generate images with improved contrast in the bright areas by shifting the intensities toward the bottom, so that bright parts are enhanced and the dark parts clipped (this will become clear in Equation (3) below). By an abuse of terminology, the authors of [8] call these images under-exposed.…”

“…In this paper, following the lines of the method introduced in [8], we describe a single image contrast enhancement method derived from EF, called Simulated Exposure Fusion (SEF). Starting from a single image, SEF simulates a bracketed exposure sequence of images and fuses them by an adaptation of EF avoiding its out-of-range artifact.…”