PerfectColor - Methods and algorithms for real-time color image enhancement

Project Summary

Color image enhancement is an important problem in many image processing applications. Development of image sensors and growing market demand for advanced solutions that are based on the image processing technology are the main factors for increasing importance of image enhancement techniques. In particular, image processing algorithms used for embedded systems such as digital cameras have strict requirements in terms of real-time operation and hardware constraints imposed by specific technology used for embedded implementation, such as FPGA technology. Most methods and algorithms published in the literature are not suitable for real-time operation due to their complexity or due to being hardware unfriendly so their embedded implementations are too complex and unable to satisfy strict requirements of real-time applications.

The overall goal of this project is to improve two important image enhancement tasks of color constancy and tone mapping found in many embedded systems such as digital cameras by proposing new practically implementable methods that are capable of real-time operation.

The main objectives of this project are research and development of:

  1. methods for real-time color constancy, and
  2. methods for real-time tone mapping.

In addition, two new image datasets that are to be created during this research, will certainly have an important impact on the whole research field of color constancy.