Difference between revisions of "Data Annotation"

From CSWiki
Jump to: navigation, search
(View Point Label)
Line 12: Line 12:
  
 
==View Point Label==
 
==View Point Label==
* Ask labelers to do a selection question with each branch to be a frame sampled from video sequences. We plan to sample those frames from equally distributed angles.
+
* Ask labelers to do a selection question with each branch to be a frame sampled from video sequences.  
 +
 
 +
We plan to sample those frames from equally distributed angles.

Revision as of 03:32, 5 January 2009

Bounding box Label

  • Annotation file format:
 Img_W Img_H xmin xmax ymin ymax
  • Problems of Current Approach
  1. some images are too large to be displayed within the screen, so have bad annotation.
  2. the "null" string in result should be replaced by -1 -1 -1 -1
  3. the following plot is that, if each image has a certain number of annotations (x axis), how many images can be inferred a true bounding box.

Valid img ratio with respect to num of annotation.png

By the plot, we suggest using 6 annotations for each image to label.

View Point Label

  • Ask labelers to do a selection question with each branch to be a frame sampled from video sequences.

We plan to sample those frames from equally distributed angles.