Difference between revisions of "Data Annotation"
From CSWiki
(→View Point Label) |
|||
| Line 2: | Line 2: | ||
* Annotation file format: | * Annotation file format: | ||
Img_W Img_H xmin xmax ymin ymax | Img_W Img_H xmin xmax ymin ymax | ||
| + | * What we have collected: | ||
| + | router | ||
| + | guitar | ||
| + | swivel chair | ||
| + | electric organ | ||
| + | * What we plan to collect: | ||
| + | "device" and "equipment" of instrumentality (instrumentation), like computer, mouse, stapler, ... | ||
| + | |||
* Problems of Current Approach | * Problems of Current Approach | ||
# some images are too large to be displayed within the screen, so have bad annotation. | # some images are too large to be displayed within the screen, so have bad annotation. | ||
Revision as of 03:47, 5 January 2009
Bounding box Label
- Annotation file format:
Img_W Img_H xmin xmax ymin ymax
- What we have collected:
router guitar swivel chair electric organ
- What we plan to collect:
"device" and "equipment" of instrumentality (instrumentation), like computer, mouse, stapler, ...
- Problems of Current Approach
- some images are too large to be displayed within the screen, so have bad annotation.
- the "null" string in result should be replaced by -1 -1 -1 -1
- 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.
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.
