The Altair Community is migrating to a new platform to provide a better experience for you. In preparation for the migration, the Altair Community is on read-only mode from October 28 - November 6, 2024. Technical support via cases will continue to work as is. For any urgent requests from Students/Faculty members, please submit the form linked here
Image Mining: Combine Global and Local Features Extraction
Legacy User
Member Posts: 0 Newbie
Dear all,
I would like to ask how one can combine the extraction of global and local features in the same workflow.
My idea was:
-Multiple Color Image Opener
-Global Feature Extractor from Single Image
-Local Feature Extractor
But or of curse this is not possible, since MCIO has only one port.
I also tried to combine 2 MCIO operators and use a join(on FileName) but this is is not working. It says that there is no FileName attribute on the on the right side (that is the Local Feature Extractor).
The Local Feature Extractor itself has the following structure(taken from some post in the forum):
-Local Feature Extractor
-Multiply
-Point of Interest
-Local Features
Thank you very much,
Elena
I would like to ask how one can combine the extraction of global and local features in the same workflow.
My idea was:
-Multiple Color Image Opener
-Global Feature Extractor from Single Image
-Local Feature Extractor
But or of curse this is not possible, since MCIO has only one port.
I also tried to combine 2 MCIO operators and use a join(on FileName) but this is is not working. It says that there is no FileName attribute on the on the right side (that is the Local Feature Extractor).
The Local Feature Extractor itself has the following structure(taken from some post in the forum):
-Local Feature Extractor
-Multiply
-Point of Interest
-Local Features
Thank you very much,
Elena
0
Answers
you can use Cartesian Product operator from Series extension to merge local features ExampleSet and global features ExampleSet. All local features from one image will have same global features.
Best,
Vaclav
in the meantime I came up with this.
I managed to combine the IMMI, DFKI and a db table in the same process.
I have no idea if this is a good practice, especially since I only used one table from the db for my test. However, this is my workflow and it is working. Your workflow with the Rename operator helped. However, in order to achieve the final result I had to use transpose twice since the DFKI operators store the path and not the FileName like IMMI, but from what I see there is no way to manipulate the entries of the example sets,only the attribute names can change. Therefore you can see the kind of ugly workaround with transpose-Rename-transpose. What do you think? Is it a good idea to have all feature extractions in one workflow (they will become more of course). Of course an easier-to-read option is saving the example sets and combine multiple retrieves I think..