Face Recognition using HANA

[Former Member]

Hi,

I am trying to gauge if HANA can be used to store bitmap images for faster face recognition as compared to RDBMS like Oracle or PostGRES. The requirement is to run a number of SQL queries against the bitmaps stored in HANA to detect a pattern match with the images being captured live via cameras. This application is intended as a security solution for public spaces. Current RDBMS is struggling with these queries and performance bottleneck needs to be removed.

Are there any metrics or case studies where HANA has been show to have an advantage for such queries, as compared to RDBMS?

Any use cases where bitmap queries have been used on HANA? I believe the data type would be BLOB.

Any help would be highly appreciated.

Binary Types

Binary types are used to store bytes of binary data.

• VARBINARY
  The VARBINARY(n) data type is used to store binary data of a specified maximum length in bytes, where n indicates the maximum length and is an integer between 1 and 5000.

Large Object (LOB) Types

LOB (large objects) data types, CLOB, NCLOB and BLOB, are used to store a large amount of data such as text documents and images. The maximum size of an LOB is 2 GB. 

• BLOB
  The BLOB data type is used to store large binary data.

[Former Member · Accepted solution]

Dear Adeel,

I haven't seen such studies, but maybe you can also ask the guys on http://www.experiencesaphana.com. However, in our HANA education (http://www.getyourhandson.it) we have loaded some PDF files into HANA as BLOB and we do a fuzzy search on this PDF documents in order to find relevant documents. I think it should work similiar for your requirement.

Best Regards,
Marcel

[rama_shankar3]

Guys:

hope also you checked out the Saphire demo on HANA demo using Xbox Kinect front-end on HANA data - if not, check it out @ http://www.youtube.com/watch?v=v6LXRNhGsHk  

Regards,

Rama