Cropping images using trained darknet model
-
class names
,confidence
andbbox coordinates
- Extract the information by running the command
./darknet detector test custom.data custom.cfg custom.weights -thresh 0.4 -dont_show -ext_output < val.txt > result-dr.txt
-
- Copy the file
result-dr.txt
to the working directory
- Copy the file
-
- Copy the image files to the folder
val-images/
- Copy the image files to the folder
-
- Remove Enter by typing
grep -vE "(Enter)" result-dr.txt > result4.txt
- Remove Detection by typing
grep -vE "(Detection)" result4.txt > result.txt
- Remove Enter by typing
-
python parsing.py