3rd year computer vision coursework - visual search system and report. Achieved 100%
Go to file
2020-02-23 19:48:47 +00:00
data done? 2019-12-03 03:30:25 +00:00
descriptor done? 2019-12-03 03:30:25 +00:00
distance working mahalanobis 2019-11-30 17:56:27 +00:00
report fixed media paths 2020-01-19 23:07:31 +00:00
util pulling results, fixed deflation 2019-12-01 23:27:29 +00:00
.gitignore updated README 2020-02-23 19:48:47 +00:00
cvpr_computedescriptors.m done? 2019-12-03 03:30:25 +00:00
cvpr_visualsearch_pca.m done? 2019-12-03 03:30:25 +00:00
cvpr_visualsearch_query_set.m done? 2019-12-03 03:30:25 +00:00
cvpr_visualsearch_rand_image.m pulling data and images 2019-12-01 12:25:26 +00:00
cvpr_visualsearch.m beginning to add results 2019-12-01 02:30:20 +00:00
labsheet3.pdf initial commit with random descriptors and skeleton report 2019-11-06 15:22:15 +00:00
parameter_iter_pca.m done? 2019-12-03 03:30:25 +00:00
parameter_iter_query_set.m done? 2019-12-03 03:30:25 +00:00
README.md updated README 2020-02-23 19:48:47 +00:00
report.pdf updated README 2020-02-23 19:48:47 +00:00
scratch.m pulling data and images 2019-12-01 12:25:26 +00:00
spec.pdf initial commit with random descriptors and skeleton report 2019-11-06 15:22:15 +00:00

Visual Search Coursework

Investigation into different methods of visual search including colour histograms, spatial colour, spatial texture and a combination of the two using MATLAB and the MSRCv2 dataset. A copy of the report can be seen here. Submitted as part of the third year EE3032 Computer Vision & Pattern Recognition module, this piece achieved full marks.

Mean average precision comparisons for different distance measures

Spatial colour mean average precision for different grid dimensions

Detected edges in one of the dataset images

Code

/data - pulled images and spreadsheet of data /descriptor - functions for extracting descriptors /distance - functions for measuring distance between descriptors /util - util functions such as toGreyscale and EVD

There are two types of script, ones that run a category response once (cvpr_visualsearch_) and ones that iteratively generate new descriptors to run queries on (parameter_)

_query_set operates using either L1 or L2 norm on the query set _pca generates an eigenmodel from the descriptors and computes mahalanobis distance _rand_image picks a random query image from each category to iterate over, no results from this script are in the paper

The cvpr_visualsearch_* scripts load descriptors from folders and perform a category response test on them.

The parameter_* scripts were used to generate iterative parameter results for descriptors. Effectively the query code from the cvpr_visualsearch_* files have been prefaced with descriptor generators that as a whole iterate over parameters instead of loading them from files.