GitHub/quartz
1
0
Fork 2
mirror of https://github.com/jackyzha0/quartz.git synced 2026-03-24 15:05:42 -05:00
Code Issues Actions Packages Projects Releases Wiki Activity

vault backup: 2023-03-31 22:16:40

Browse Source
This commit is contained in:
Jet Hughes 2023-03-31 22:16:40 +13:00
parent 6856ff7704
commit 3431d3185a
1 changed files with 11 additions and 14 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

25
content/notes/342-assignment-01.md
View File

@ -15,33 +15,28 @@ Choosing images
use a range of input images use a range of input images
# Experiment 1: Feature Matching # Experiment 1: Feature Matching
Highly detailed images will have a lot of information such as trees, text, buildings, landscapes etc. Less detailed images with be sparse, these could be bare buildings, walls, images with large blocks of color, etc. Detailed Images will result in a large number of features which are tightly grouped. I predict this will reduce the accuracy of FLANN based matching.
HYPOTHESIS: HYPOTHESIS:
Level of detail in images will have a significant effect on the accuracy of FLANN based matching FLANN based matching will decrease in accuracy as level of detail increases
EXPERIMENT DESIGN: EXPERIMENT DESIGN:
Variables Variables
- Independent Variables: Level of detail (Number of SIFT features detected) - Independent Variables: Level of detail (Number of SIFT features detected)
- Dependent Variable: reprojection error of image mosaics - Dependent Variable: Reprojection error of feature matches
Collect a number of highly detailed pairs of images and the same number pairs of less detailed images. Highly detailed images will have a lot of information such as trees, text, buildings, landscapes etc. Less detailed images with be sparse, these could be bare buildings, walls, images with large blocks of color, etc. Procedure
1. Select a set of image pairs with various levels of detail
Detailed Images will result in a large number of features which are tightly grouped. This could significantly affect the accuracy of FLANN based matching
To conduct the experiment I will measure the reprojection error of image mosiacs created using FLANN and Brute force matching
1. Select a set of highly detailed image pairs, and a set of sparsly detailed image pairs
2. For each image pair, detect SIFT features and perform both FLANN and Brute force matching 2. For each image pair, detect SIFT features and perform both FLANN and Brute force matching
3. Calculate the reprojection error of features matched using both methods for each image pair 3. Calculate the reprojection error of features matched using both methods for each image pair
4. For each method, plot the accuracy as a function of the number of features/level of detail 4. For each method, plot the accuracy as a function of the number of features/level of detail
_control for other factors: lighting, camera settings (exposure, iso, etc), resolution?__ _control for other factors: lighting, camera settings (exposure, iso, etc), resolution?__
**hypothesis super wordy **hypothesis super wordy**: fixed
what is your error metric **what is your error metric**: reprojection error
how will you decide statistically if your hypothesis is true or not** **how will you decide statistically if your hypothesis is true or not**: measure correlation between number of features and level of detail
RESULTS RESULTS
@ -53,6 +48,8 @@ QUESTION: How does the choice of RANSAC threshold affect the speed of homography
HYPOTHESIS: The choice of RANSAC threshold affects the speed of homography estimation in highly detailed images. HYPOTHESIS: The choice of RANSAC threshold affects the speed of homography estimation in highly detailed images.
As the RANSAC threshold increases the speed of homography estimation decreases
EXPERIMENT DESIGN: EXPERIMENT DESIGN:
Variables Variables