Hybrid Images

Side coursework for COMP6223 (computer vision) based on the paper of Oliva et. al. 2006. The report refers to two methods: a spatial approach and a frequency approach.

master branch:

Contains all necessary files to run the spatial approach.

• data: contains all image pairs
• report.pdf: report
• my_conv.m: implements a convolution operation
• main_spatial.m: Runs main file. Run this for the spatial approach
• lowpass.m/highpass.m: lowpass & highpass filters based on conv. operations
• spatial_representation.m: sublots of enhanced images in the spatial domain
• freq_representation.m: subplots of enhanced images in the frequency domain

freq-method branch:

Contains all necessary files to run the frequency approach.

• main_frequency.m: Runs main file. Run this for the frequency approach
• LP_fourier.m/HP_fourier.m: lowpass & highpass filters based on element wise operations on the FFT of images

BibTex entry of Oliva et. al. 2006:

 @article{Oliva:2006:HI:1141911.1141919,
 author = {Oliva, Aude and Torralba, Antonio and Schyns, Philippe G.},
 title = {Hybrid Images},
 journal = {ACM Trans. Graph.},
 issue_date = {July 2006},
 volume = {25},
 number = {3},
 month = jul,
 year = {2006},
 issn = {0730-0301},
 pages = {527--532},
 numpages = {6},
 url = {http://doi.acm.org/10.1145/1141911.1141919},
 doi = {10.1145/1141911.1141919},
 acmid = {1141919},
 publisher = {ACM},
 address = {New York, NY, USA},
 keywords = {human perception, hybrid images, scale space},
 }