Title: Digital Album Overview
1Digital Album Overview
Launch Application
Update Tree Hierarchy
Database
Import Images
For Each Event Find Related Events Using
WordNet Associations Based On User Annotation
Manually
Parse/Read Image Metadata Lat/Long/Data/Filename
Update Database with WordNet Keywords using User
Event Annotation as Driver
User Event Annotation
Parse/Read Update Named Location
Update Database with DateTime Events
Update Events With Location
Display Thumbnail Images Place Marker on Map
Sort by Date and Segment using a
DateTime Sliding Window Algorithm
Clustering
Import
Scott Bishop San Francisco State university,
December 6, 2005
2Image MetaData, Map Display
- Digital Images meta-data extraction
- Scanned in images then manually tagged with
Lat/Long and datetime - Digital camera images then manually tagged
- Exif Headers parsed, data extracted and imported
to MySQL Database
- Map Display
- Lat/Long of images passed to MapPanel and
displayed on the map. Marker size is scaled
logarithmically to the number of images in the
database. - Marker Size klog10 (N)2 where
- k constant N number of Images
- and 2 Default Size
Scott Bishop San Francisco State university,
December 6, 2005
3Image and Event Location Names
Image Locations Bounding Box Image location
detection (reverse geo-coding). Search a cross
reference table until a City name is found and
use it for the image location name
Image_1
Example In a boat offshore of San Francisco,
what name should be used? Pacific Ocean? Daly
City? San Francisco?
Event Locations Event Names Concatenate each
distinct images location name in the events
and use in the Hierarchy
Scott Bishop San Francisco State university,
December 6,, 2005