A research poster titled "Zero-Reference Low-Light Enhancement" is displayed, featuring contributions from Wenjing Wang, Huan Yang, and Jianlong Fu. It is showcased at the Computer Vision and Pattern Recognition (CVPR) conference in Seattle, WA. The poster is divided into several sections, including "Introduction" and "Method." 

The "Introduction" describes the objective to enhance low-light images without requiring reference images. The aim is to improve robustness, data usage during training, and illumination quality under poor conditions.

The "Method" section details the training framework, which utilizes a physical quadruple paired constraint and reconstructs the prior back to illumination. Diagrams and images illustrate the training and testing processes, showing how the method bridges the gap between normal and low-light images. Solutions to detail degradation and the effects of the enhancement are also depicted through comparative visuals.

In the background, an individual in gray pants and brown shoes stands near the poster, partially visible at the bottom edge of the image.
Text transcribed from the image:
CVPR Zero-Reference Low-L
SEATTLE
Introduction
Wenjing Wang Huan Yang Jianlong
Zero-erence Low-Light Enhancement lam
solely with nomalight images aducing the read
Method
Training framework
Pack a pryce quadruple prio
construct the prior back to imec
Our aim improve the bes
Octa usage ouring hig
umination specific hyper
FGH
Our methodology dragan Bumisation
Solution of detal degraddon bypass