Unsupervised Deep Contrast Enhancement With Power Constraint for OLED Displays

Abstract

Various power saving and contrast enhancement (PSCE) techniques have been applied to an organic light emitting diode (OLED) display for reducing the power demands of the display while preserving the image quality. In this paper, we propose a new deep learning-based PSCE scheme that can save power consumed by the OLED display while enhancing the contrast of the displayed image. In the proposed method, the power consumption is saved by simply reducing the brightness a certain ratio, whereas the perceived visual quality is preserved as much as possible by enhancing the contrast of the image using a convolutional neural network (CNN). Furthermore, our CNN can learn the PSCE technique without a reference image by unsupervised learning. Experimental results show that the proposed method is superior to conventional ones in terms of image quality assessment metrics such as a visual saliency-induced index (VSI) and a measure of enhancement (EME). 1 Index Terms-Convolutional neural network, deep learning, energy efficiency, image enhancement.

Technologies based on energy savings for OLED devices

“…Deep learning outperforms the other methods in terms of image enhancements [17,18]. It is used to enhance power consumption by reducing it while viewing the most improved photo in terms of brightness and details on an OLED display [15]. This display's total dissipated power (TDP) that has N pixels is expressed according to [4,19,20] as follows:…”

“…For the loss function, training without the reference image was made as it is related to unsupervised learning. By doing this, the enhancement of image's contrast is applied, and at the same time, power consumption is being saved [15]. The method divides the image into two images, luminance, and color.…”

“…The method divides the image into two images, luminance, and color. The luminance's intensity reduces to decrease the power demand, of course, the contrast in this way improved by using the deep learning method, CNN, and after that, the resulted photo will be added to the color image as shown in Figure 1 [15]. After reducing the luminance of the input image, the rate of power saving (R) will be reduced as well as shown in Eq.…”

“…Pdim and Pin are TDP of dimmed and input images, and k is a ratio of reducing the luminance image's intensity. By implementing this, the image quality may reduce and lose its details, for that the adaptive contrast enhancement (ACE) was introduced to improve the contrast of darkening image [15].…”

“…Some methods were implemented to reduce power consumption while keeping and preserving the details of images displayed on OLED display. One of the methods is depending on the power-constrained contrast enhancement (PCCE) implemented by using one of deep learning algorithms [15]. Another method is by darkening the image by reducing the luminance of some pixels according to the displayed image on OLED display, in other words, remapping the colors of the image [8].…”