Paintings / Photos geolocalisation: Difference between revisions
Junhong.li (talk | contribs) (→SIFT) |
Junhong.li (talk | contribs) (→SIFT) |
||
| Line 41: | Line 41: | ||
* Dataset spliting | * Dataset spliting | ||
To check the feasibility of our method, we try to use images with geo-coordinates to test. Therefore, we should split our dataset. In our experiment, because we do not have a dataset large enough and the matching without parallel is time-consuming, we randomly choose 2% of the dataset to be test dataset. | To check the feasibility of our method, we try to use images with geo-coordinates to test. Therefore, we should split our dataset. In our experiment, because we do not have a dataset large enough and the matching without parallel is time-consuming, we randomly choose 2% of the dataset to be test dataset. | ||
* Scale-invariant feature detection | |||
* Keypoints matching | |||
* Error analysis | |||
For each match-pair, we calulate the MSE(mean square error) of lantitude and longitude. In order to assess our result, We try to visualise the distribution of MSE and give a 95% CI of median value of MAE by bootstrapping. | |||
===Deep Learning=== | ===Deep Learning=== | ||
Revision as of 12:56, 9 December 2020
Introduction
Planning
| Date | Task |
|---|---|
| By 08.Nov. | Determine the method to be used and obtain the dataset |
| By 14.Nov. | Get result of SIFT method and evaluate the result |
| By 22.Nov. | Use deep learning methods to obtain preliminary results |
| By 29.Nov. | Finalize result of deep learning and try improvements |
| By 06.Dec. | Implement web using Django framework |
| By 10.Dec. | Final report writing |
| By 15.Dec. | Final result and presentation |
Methodology
Data collection
SIFT
SIFT, scale-invariant feature transform, is a feature detection algorithm to detect and describe local features in images. We try to use this method to detect and describe key points in the image to be geolocalised and images with geo-coordinates. With these key points, we can find the most similar image and then finish the geolocalisation.
- Dataset spliting
To check the feasibility of our method, we try to use images with geo-coordinates to test. Therefore, we should split our dataset. In our experiment, because we do not have a dataset large enough and the matching without parallel is time-consuming, we randomly choose 2% of the dataset to be test dataset.
- Scale-invariant feature detection
- Keypoints matching
- Error analysis
For each match-pair, we calulate the MSE(mean square error) of lantitude and longitude. In order to assess our result, We try to visualise the distribution of MSE and give a 95% CI of median value of MAE by bootstrapping.