Coventry University Featured PhD Programmes
University of Kent Featured PhD Programmes
University of Huddersfield Featured PhD Programmes
Catalysis Hub Featured PhD Programmes
University of Reading Featured PhD Programmes

Image colour transfer and AI: new algorithms for estimating colour transfer between images and video frames (GONGHU19SCIC)

This project is no longer listed on FindAPhD.com and may not be available.

Click here to search FindAPhD.com for PhD studentship opportunities
  • Full or part time
    Dr H Gong
    Dr M Mackiewicz
  • Application Deadline
    No more applications being accepted
  • Competition Funded PhD Project (Students Worldwide)
    Competition Funded PhD Project (Students Worldwide)

Project Description

Colours may vary from image to image. This may be caused by different capture conditions (e.g. illumination, camera), seasons, object reflectances, and post-processing pipelines. Colour transfer is a process that transfers one image’s colour palette to another image. While there are some existing solutions, they often require a large number of modelling parameters [iv], high computational cost [i], or expensive image registration [iii]. Without the loss of accuracy, it is often desirable to avoid complex image registrations and artefacts while unifying the colours among different captures.

In this PhD project, we aim to develop new algorithms for estimating the colour transfer between two images or among a set of video frames. These potential algorithms provide powerful applications in particular the film industry (e.g. “Match Color” in Apple FinalCut) whereby colour transfer estimation is often required for blending one video sequence with a different video sequence or some generated visual effects. In mobile device video capturing apps, a video colour stabilization [i-iii] is also desirable for providing a smooth and consistent colour transition between video frames. The project will begin by inspecting classic methods for colour transfer estimation based on image intensity statistics [i-iii]. We then combine the classic approaches with the deep neural networks [iv,v] to achieve a real-time and high-quality performance. Depending on the application scenario, we plan to develop two different solutions of colour transfer for the same scene (useful for video colour stabilisation) and the unknown different scenes (useful for creative graphics editing).

For more information on the supervisor for this project, please go here: https://people.uea.ac.uk/en/persons/h-gong

Type of programme: PhD

Project start date: October 2019

Mode of study: Full time

Entry requirements: Acceptable first degree - Computer Science, Mathematics, Physics.
The standard minimum entry requirement is 2:1.



Funding Notes

This PhD project is in a Faculty of Science competition for funded studentships. These studentships are funded for 3 years and comprise home/EU fees, an annual stipend of £14,777 and £1,000 per annum to support research training. Overseas applicants may apply but they are required to fund the difference between home/EU and overseas tuition fees (which for 2018-19 are detailed on the University’s fees pages at https://portal.uea.ac.uk/planningoffice/tuition-fees . Please note tuition fees are subject to an annual increase).

References

i) Farbman, Zeev, and Dani Lischinski. "Tonal stabilization of video." ACM Transactions on Graphics (TOG). Vol. 30. No. 4. ACM, 2011.

ii) Frigo, Oriel, et al. "Motion driven tonal stabilization." IEEE Transactions on Image Processing 25.11 (2016): 5455-5468.

iii) Gong, Han, et al. "3D color homography model for photo-realistic color transfer re-coding." The Visual Computer (2017): 1-11.

iv) Luan, Fujun, et al. "Deep Photo Style Transfer." 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). IEEE, 2017.

v) Liu, Ming-Yu, Thomas Breuel, and Jan Kautz. "Unsupervised image-to-image translation networks." Advances in Neural Information Processing Systems. 2017.



FindAPhD. Copyright 2005-2019
All rights reserved.