Name: TattDL
Owner: NASA JPL MEMEX
Owner: NASA JPL MEMEX
Description: Tattoo detection and localization
Forked from: z-harry-sun/TattDL
Created: 2016-07-22 15:08:50.0
Updated: 2018-02-25 22:59:51.0
Pushed: 2017-07-13 23:08:38.0
Homepage: null
Size: 601
Language: Python
GitHub Committers
User | Most Recent Commit | # Commits |
---|
Other Committers
User | Most Recent Commit | # Commits |
---|
Tattoo detection and localization
This uses the Caffe hash: 4115385deb3b907fcd428ac0ab53b694d741a3c4
The tattoo model is managed by git-lfs, so make sure that is installed (see
https://git-lfs.github.com/
for details on how to do that).
Then, fetch the file using:
git lfs fetch git lfs checkout
After running the `tool, one of the output files will be a
detection.txtfile.
This is a CSV-like file (
|` primary separators):
filename | proc time | scale | scores | boxes
The boxes field (the last one) may container 0 or more confidence and bounding box specifications. Separating this text field by spaces () will yield one or more sub-CSV rows with the format:
confidence,x,y,width,height
X and Y coordinates specify the upper left corner of the sub-region, assuming (0,0) is the upper left corner of the image.
Please cite the following paper if you use this software in your work:
roceedings{sun2016tattoo,
tle={Tattoo detection and localization using region-based deep learning},
thor={Sun, Zhaohui H and Baumes, Jeff and Tunison, Paul and Turek, Matt and Hoogs, Anthony},
oktitle={Pattern Recognition (ICPR), 2016 23rd International Conference on},
ges={3055--3060},
ar={2016},
ganization={IEEE}
Additionally, citing the JPL MEMEX project `http://memex.jpl.nasa.gov/
` would be welcomed.
Tattoo detection and localization
This uses the Caffe hash: 4115385deb3b907fcd428ac0ab53b694d741a3c4
The tattoo model is managed by git-lfs, so make sure that is installed (see
https://git-lfs.github.com/
for details on how to do that).
Then, fetch the file using:
git lfs fetch git lfs checkout
After running the `tool, one of the output files will be a
detection.txtfile.
This is a CSV-like file (
|` primary separators):
filename | proc time | scale | scores | boxes
The boxes field (the last one) may container 0 or more confidence and bounding box specifications. Separating this text field by spaces () will yield one or more sub-CSV rows with the format:
confidence,x,y,width,height
X and Y coordinates specify the upper left corner of the sub-region, assuming (0,0) is the upper left corner of the image.
Please cite the following paper if you use this software in your work:
roceedings{sun2016tattoo,
tle={Tattoo detection and localization using region-based deep learning},
thor={Sun, Zhaohui H and Baumes, Jeff and Tunison, Paul and Turek, Matt and Hoogs, Anthony},
oktitle={Pattern Recognition (ICPR), 2016 23rd International Conference on},
ges={3055--3060},
ar={2016},
ganization={IEEE}
Additionally, citing the JPL MEMEX project `http://memex.jpl.nasa.gov/
` would be welcomed.