Autoexposure (AE) is a critical step applied by camera systems to ensure properly ex- posed images. While current AE algorithms are effective in well-light environments with unchanging illumination, these algorithms still struggle in environments with bright light sources or scenes with abrupt changes in lighting. A significant hurdle in developing new AE algorithms for challenging environments, especially those with time-varying lighting, is the lack of platforms to evaluate and improve AE algorithms and suitable image datasets. To address this issue, we have designed a software platform to allow AE algorithms to be used in a plug-and-play manner with the dataset. In addition, we have captured a new 4D exposure dataset that provides a complete solution space (i.e., all possible exposures) over a temporal sequence with moving objects, bright lights, and varying lighting. Our dataset and associate platform enable repeatable evaluation of different AE algorithms and provide a much-needed starting point to develop better AE methods. We examine several existing AE strategies using our dataset and show that users prefer a simple saliency method for challenging lighting conditions.

The 4D Dataset & platform

We produced a high resolusion 4 dimention RAW image dataset with 36000 images in a size of 850G. Where 4D refers to Time, Exposure ( shutter speed ), Width and Height.

The dataset allows AE algorithms to be tested repeatedly in a time step manner, the right side image shows results of 4 AE methods on a same time stamp of this scene.

A software platform is designed to evaluate AE algorithms and to visualize the reuslts. The image illustrates basic steps for using our AE platform. The user selects a scene and an algorithm. Parameters of the AE algorithm can adjust. After the AE algorithm runs, the platform plays the output images and the corresponding plots at 10 FPS. The user cause ``pause'' at any time frame to adjust the exposure stack slider for comparison. Image histogram for each frame are also shown.

It is hopeful that our new dataset and platform will advance the future Autoexposure Algorithms.