Description: Retrieving and segmenting your magazine within one snapshot
Scenario:
1. User takes a snapshot on a paper document (e.g. magazine)
2. Retrieve the corresponding digital document (pdf file)
3. System can auto-determine the reading sequence, layout, and segmentation for user
We defined two parts to work on (a short-term goal)
a. 'matching part'
given a camera image, return the matched ground truth image
b. 'segmentation part'
given a document, return the mobile-readable text blocks and image blocks
Retrieving and segmenting your magazine within one snapshot



