convert2bo's Introduction

Convert2Yolo

Object Detection & Recognition DataSet annotation Convert Yolo Darknet

Pre-Requiredment

1. Make Directory

All convert code need some directory hierarchy

just make folder like this


[some your project root directory]
|
| - [JPEG] : For images folder
|
| - [label] : For annotaion folder
|
| - [results] : For results of convert2Yolo

2. Make Classe file

you should make class file

follow this format

reference voc.names file in repository

voc.names

aeroplane
bicycle
bird
boat
bottle
bus
car
cat
chair
cow
diningtable
dog
horse
motorbike
person
pottedplant
sheep
sofa
train
tvmonitor

3. Install requirement package

pip3 install -r pip_install_required.txt

VOC Pascal Format

Here is VOC Pascal DataSet example

1. example Code

python voc.py test.names example/voc/JPEG/ example/voc/label/ example/voc/results/

TODO

Support VOC Pascal Format
Support Udacity Format
Write how to using VOC Pascal Format in ReadMe
Write how to using Udacity Format in ReadMe
Code Refactoring

convert2bo's People

Stargazers

Watchers

convert2bo's Issues

Does not support polygon tag

This software does not seem to support the polygon tag and works only with the bndbox tag

Thus, it is basically useless, as it is better to use a polygon than a square to tag an object since it does not carry so much junk around the edges

ERROR : invalid literal for int() with base 10: '362.2251767966361'

Changed lines 427 to 430 of convert2Yolo.py to this instead which fixed the problem.

xmin = float(bndbox.find('xmin').text)
xmax = float(bndbox.find('xmax').text)
ymin = float(bndbox.find('ymin').text)
ymax = float(bndbox.find('ymax').text)

Recommend Projects