I. To locate the most valuable matches as the most important & correlated relationship
- i.e. to keep the matches much more concentrated
Thursday, 24 April 2014
Wednesday, 23 April 2014
Work of Camera Calibration (11)
I. Check every parameters in order to increase the number of MSER features
- 'ThresholdDelta': decrease to 0.2
- 'MaxAreaVariation': increase to 0.85
- 'RegionAreaRange': change to [5 sizeOfImage*30%]
II. Check every parameters in order to increase the number of SURF features (in 128-bit)
- 'MetricThreshold': decrease to 100
- 'NumScaleLevels': increase to 6
- 'NumOctaves': increase to 4
III. Check the 2 function calls to know what value is going to be feedback
- 'extractFeatures': scale, signOfLaplacian, orientation, location, metric, count
- 'detectMSERFeatures': count, location, axes, orientation, pixelList
IV. Check every parameters in 'Matching Function Call' in order to increase the number of matches
- 'Method': change to 'NearestNeighborSymmetric'
- 'MatchThreshold': increase to 10~20
- 'Metric': change to 'SAD'
- 'MaxRatio': 'Default'
Finally, we have 400~600 'MSER+SURF(in 128-bit)' features and 300~450 matches.
- 'ThresholdDelta': decrease to 0.2
- 'MaxAreaVariation': increase to 0.85
- 'RegionAreaRange': change to [5 sizeOfImage*30%]
II. Check every parameters in order to increase the number of SURF features (in 128-bit)
- 'MetricThreshold': decrease to 100
- 'NumScaleLevels': increase to 6
- 'NumOctaves': increase to 4
III. Check the 2 function calls to know what value is going to be feedback
- 'extractFeatures': scale, signOfLaplacian, orientation, location, metric, count
- 'detectMSERFeatures': count, location, axes, orientation, pixelList
IV. Check every parameters in 'Matching Function Call' in order to increase the number of matches
- 'Method': change to 'NearestNeighborSymmetric'
- 'MatchThreshold': increase to 10~20
- 'Metric': change to 'SAD'
- 'MaxRatio': 'Default'
Finally, we have 400~600 'MSER+SURF(in 128-bit)' features and 300~450 matches.
Tuesday, 22 April 2014
Summary of Weekly Meeting - (7)
1. Preparing Powerpoint for Weekly meeting; and trying to make a conclusion based on my current result; then giving a possible solution.
2. Summary of Weekly meeting:
- Paying too much effort to achieve the final goal -> should substitute by using direct way
- Giving a final tune of image matching in order to know I have done exhausted way to deal with that; then, we can go to the next phase in the individual project
- Currently just stop for considering 3D matching stuff as well as build up a GUI for final presentation: don't go too far while doesn't close current step which has done an exhausted process for confirming that there is no way to improve the proposed method anymore.
2. Summary of Weekly meeting:
- Paying too much effort to achieve the final goal -> should substitute by using direct way
- Giving a final tune of image matching in order to know I have done exhausted way to deal with that; then, we can go to the next phase in the individual project
- Currently just stop for considering 3D matching stuff as well as build up a GUI for final presentation: don't go too far while doesn't close current step which has done an exhausted process for confirming that there is no way to improve the proposed method anymore.
Work of Camera Calibration (10)
1. Applying my work on our collecting dataset
2. Editing image stitching code (...maintaining in 2D space)
2. Editing image stitching code (...maintaining in 2D space)
Monday, 14 April 2014
Work of Camera Calibration (9)
1. Testing in Released Datasets
2. Checking my codes and results to see whether my work is good or not while applying on the officially released datasets.
2. Checking my codes and results to see whether my work is good or not while applying on the officially released datasets.
Friday, 11 April 2014
Work of Camera Calibration (8)
I. Work:
。Find proposed source code of matching by using the Affine Invariant descriptor
。Paper surveys
II. Learning how to create a mechanism of change the perspective of an image, and generating some relevant values of parameters for future matching process
Thursday, 10 April 2014
Work of Camera Calibration (8) : Meeting and Having discussion with Matina (2)
I. Using one of the affine invariant descriptors (MSER + SURF)
。Having better result than just using SIFT descriptor, but need to be tuned because of having only few
features which the number is less than a hundred
II. Summary of discussion
。Need to try every well-known affine invariant descriptor (especially using the proposed source code) to
find which method can provided better matching results in our datasets.
。Do more literature review & find source code of method which are using these affine invariant descriptors
。The essence of this testing experiment (or testing) is finding the method which does well in rotation,
translation, scaling (...affine transformation is not a serious problem in our work)
。For each source code (or/and the affine invariant descriptors), need to know when it does well, why it
works, and how it works (especially pay attention to the size of the descriptor / the measurement of
similarity / why parts of content in the frame doesn't show any features such as the corner of cabinet)
。Having better result than just using SIFT descriptor, but need to be tuned because of having only few
features which the number is less than a hundred
II. Summary of discussion
。Need to try every well-known affine invariant descriptor (especially using the proposed source code) to
find which method can provided better matching results in our datasets.
。Do more literature review & find source code of method which are using these affine invariant descriptors
。The essence of this testing experiment (or testing) is finding the method which does well in rotation,
translation, scaling (...affine transformation is not a serious problem in our work)
。For each source code (or/and the affine invariant descriptors), need to know when it does well, why it
works, and how it works (especially pay attention to the size of the descriptor / the measurement of
similarity / why parts of content in the frame doesn't show any features such as the corner of cabinet)
Subscribe to:
Posts (Atom)