Description
In this project, we implement an end-to-end pipeline to learn user preferences to enhance images in a personalized way.
The five major components of this project are: computing a distance metric, finding a training set that maximally represents the dataset, finding an optimal parameter set for each training image, training, and finally, enhancing the images.
The efficiency of this approach lies in the fact that almost all of the processing can be done offline, so the user is involved only in a short training phase. To test the validity of this method, we carried out user studies, and the fact that our method was preferred over professional software for some images shows the potential of this approach.
Image Processing Pipeline to Personally Enhance Newly Added Image
Training GUI implemented in MATLAB
https://web.stanford.edu/class/ee368/Project_Autumn_1617/Reports/report_shah_luppescu.pdf
Reviews
There are no reviews yet.