Collaborative memory organization
All family members can help to correct the stories of
photos collaboratively. Since our story contents were
collected from transcribing the voice recordings by ASR
services, some words may not be correctly transcribed,
especially names of people or locations. We supposed
that the family members should know the exact story,
so we implemented an editing mode for the members
to re-write or summarize the voice recordings and text
messages, the interface shows in Figure 2. Meanwhile,
we extracted meaningful tags from the corrected story
contents. For tags extraction, we adopt the CKIP text
segmentation system, which performs as a web service
and returns segmented text with part-of-speech
tagging. Our system removed the stop words and
annotated the nouns onto the photos as tags, which
can be used to describe and highlight particular objects
in the photos. Furthermore, the tags also can be
accustomed to organizing and retrieving related photos.