goodhamgupta/yolo_to_voc.py

isennkubilay · 2019-09-27T11:42:57Z

Which part of your code do I need to change to implement on my code?
For example: CLASS_MAPPING = {
'0': 'horse' I change to my class, ANNOTATIONS_DIR_PREFIX and DESTINATION_DIR
What do I need to do more?

Usmansagemode · 2019-10-21T13:15:34Z

@joskaaaa

i have this error any idea what is happening here . why program is looking for the image at this location ?

Did you find a way out of this problem? I am getting the same error. it doesnt make sense

AdonaiVera · 2019-10-31T03:46:17Z

Hey, this is a second version of the solution, if you wanna use it you have to change the path of the image and txt file.
Besides, we have image and txt file in the same folder and we chake if the txt file is empty beucase it may cause error.

# Script to convert yolo annotations to voc format
import os
import xml.etree.cElementTree as ET
from PIL import Image

ANNOTATIONS_DIR_PREFIX = "./Database1"

DESTINATION_DIR = "./DataBaseXML"

CLASS_MAPPING = {
    '0': 'drone'
    # Add your remaining classes here.
}


def create_root(file_prefix, width, height):
    root = ET.Element("annotations")
    ET.SubElement(root, "filename").text = "{}.JPEG".format(file_prefix)
    ET.SubElement(root, "folder").text = "Database1"
    size = ET.SubElement(root, "size")
    ET.SubElement(size, "width").text = str(width)
    ET.SubElement(size, "height").text = str(height)
    ET.SubElement(size, "depth").text = "3"
    return root


def create_object_annotation(root, voc_labels):
    for voc_label in voc_labels:
        obj = ET.SubElement(root, "object")
        ET.SubElement(obj, "name").text = voc_label[0]
        ET.SubElement(obj, "pose").text = "Unspecified"
        ET.SubElement(obj, "truncated").text = str(0)
        ET.SubElement(obj, "difficult").text = str(0)
        bbox = ET.SubElement(obj, "bndbox")
        ET.SubElement(bbox, "xmin").text = str(voc_label[1])
        ET.SubElement(bbox, "ymin").text = str(voc_label[2])
        ET.SubElement(bbox, "xmax").text = str(voc_label[3])
        ET.SubElement(bbox, "ymax").text = str(voc_label[4])
    return root


def create_file(file_prefix, width, height, voc_labels):
    root = create_root(file_prefix, width, height)
    root = create_object_annotation(root, voc_labels)
    tree = ET.ElementTree(root)
    

    tree.write("{}/{}.xml".format(DESTINATION_DIR, file_prefix))


def read_file(file_path):
    file_prefix = file_path.split(".txt")[0]
    image_file_name = "{}.JPEG".format(file_prefix)
    img = Image.open("{}/{}".format("Database1", image_file_name))
    print(img)

    w, h = img.size
    prueba = "{}/{}".format("Database1", file_path)
    print(prueba)
    with open(prueba) as file:
        lines = file.readlines()
        voc_labels = []
        for line in lines:	
            voc = []
            line = line.strip()
            data = line.split()
            voc.append(CLASS_MAPPING.get(data[0]))
            bbox_width = float(data[3]) * w
            bbox_height = float(data[4]) * h
            center_x = float(data[1]) * w
            center_y = float(data[2]) * h
            voc.append(center_x - (bbox_width / 2))
            voc.append(center_y - (bbox_height / 2))
            voc.append(center_x + (bbox_width / 2))
            voc.append(center_y + (bbox_height / 2))
            voc_labels.append(voc)
        create_file(file_prefix, w, h, voc_labels)
    print("Processing complete for file: {}".format(file_path))


def start():
    if not os.path.exists(DESTINATION_DIR):
        os.makedirs(DESTINATION_DIR)
    for filename in os.listdir(ANNOTATIONS_DIR_PREFIX):
        if filename.endswith('txt'):
            try:
                PathFileName = "{}/{}".format("Database1", filename)
                if os.stat(PathFileName).st_size > 0:
                    print("Si")
                    read_file(filename) 
            except:
                print("No")         
            
        else:
            print("Skipping file: {}".format(filename))


if __name__ == "__main__":
    start()

`

devika-28 · 2019-11-07T14:04:21Z

List Index Out Of Range Error

devika-28 · 2019-11-07T14:59:18Z

List Index Out Of Range Error

the tree consists of the following data, but I am not getting the desired output

Rutvikshah47 · 2019-11-14T12:17:38Z

It is working. Thanks @AdonaiVera.

HiddenFox29 · 2019-11-20T09:54:53Z

Hi people! Pay attention to the image extension. In my case, there was an extension .jpg (lower case), and I had to replace .JPEG so that the script worked fine.

khalidw · 2019-12-09T15:05:43Z

A useful tip:

Make sure that you have all your images in a folder name images in the same directory where you execute this script

I would like to suggest one edit on line #78
Change
with open(file_path, 'r') as file:
To
with open(os.path.join(ANNOTATIONS_DIR_PREFIX,file_path), 'r') as file:

Otherwise script generates an error, as it tries to read the filename (label .txt file) without complete path.

SiBensberg · 2020-01-21T14:26:09Z

@khalidw
thank you very much. Helped a lot.
@devika-28
Added a "if" in line 85. So the "List Index Out Of Range Error" is fixed

if len(data) == 0: continue;

ashwath1295 · 2020-03-24T18:46:19Z

A useful tip: this is the script to change the file names from jpg to .JPEG

import os, sys

dir =("C:\Users\ashwa\Downloads\Downloads\visiolab-food-detection\train\labels")
for i in os.listdir(dir):
files = os.path.join(dir,i)
split= os.path.splitext(files)
if split[1]=='.jpg':
os.rename(files,split[0]+'.JPEG')

ANNOTATIONS_DIR_PREFIX = "./Database1"
and change the Database1 name in the program with the respective folder name

vikashkrs · 2020-04-02T14:17:37Z

HI guys,

I am trying to convert my yolo annotated file with this script but its not working. please let me know where is the issue .

i have provided my label txt file path as input to this parameter ANNOTATIONS_DIR_PREFIX

mjahanifar · 2020-04-09T13:41:54Z

Hey guys, I edited the code a little so you can apply it to your data. Jut set directories and class names as you wish:

# Script to convert yolo annotations to voc format
import os
import xml.etree.cElementTree as ET
from PIL import Image
from math import floor

###########################
#set directories and class names
###########################
ANNOTATIONS_DIR_PREFIX = "D:/DeepData/DeepHemato/Data/Train/yoloAnots"
IMAGE_DIR_PREFIX = "D:/DeepData/DeepHemato/Data/Train/images"
imgExt = "png"
imgChnls = 3 #RGB:3 ; Grayscale:1

DESTINATION_DIR = "D:/DeepData/DeepHemato/Data/Train/vocAnots"

CLASS_MAPPING = {
    '0': 'Neutrophil',
    '1': 'Eosinophil',
    '2': 'Basophil',
    '3': 'Lymphocyte',
    '4': 'Monocyte',
    '5': 'Denature'
}

###########################

def create_root(file_prefix, width, height):
    root = ET.Element("annotations")
    ET.SubElement(root, "filename").text = "{}.{}".format(file_prefix,imgExt)
    ET.SubElement(root, "folder").text = "{}/{}.{}".format(IMAGE_DIR_PREFIX,file_prefix,imgExt)
    size = ET.SubElement(root, "size")
    ET.SubElement(size, "width").text = str(width)
    ET.SubElement(size, "height").text = str(height)
    ET.SubElement(size, "depth").text = str(imgChnls)
    return root


def create_object_annotation(root, voc_labels):
    for voc_label in voc_labels:
        obj = ET.SubElement(root, "object")
        ET.SubElement(obj, "name").text = voc_label[0]
        ET.SubElement(obj, "pose").text = "Unspecified"
        ET.SubElement(obj, "truncated").text = str(0)
        ET.SubElement(obj, "difficult").text = str(0)
        bbox = ET.SubElement(obj, "bndbox")
        ET.SubElement(bbox, "xmin").text = str(voc_label[1])
        ET.SubElement(bbox, "ymin").text = str(voc_label[2])
        ET.SubElement(bbox, "xmax").text = str(voc_label[3])
        ET.SubElement(bbox, "ymax").text = str(voc_label[4])
    return root


def create_file(file_prefix, width, height, voc_labels):
    root = create_root(file_prefix, width, height)
    root = create_object_annotation(root, voc_labels)
    tree = ET.ElementTree(root)
    

    tree.write("{}/{}.xml".format(DESTINATION_DIR, file_prefix))


def read_file(file_path):
    file_prefix = file_path.split(".txt")[0]
    image_file_name = "{}.{}".format(file_prefix,imgExt)
    img = Image.open("{}/{}".format(IMAGE_DIR_PREFIX, image_file_name))
    print(img)

    w, h = img.size
    prueba = "{}/{}".format(ANNOTATIONS_DIR_PREFIX, file_path)
    print(prueba)
    with open(prueba) as file:
        lines = file.readlines()
        voc_labels = []
        for line in lines:	
            voc = []
            line = line.strip()
            data = line.split()
            voc.append(CLASS_MAPPING.get(data[0]))
            bbox_width = float(data[3]) * w
            bbox_height = float(data[4]) * h
            center_x = float(data[1]) * w
            center_y = float(data[2]) * h
            voc.append(floor(center_x - (bbox_width / 2)))
            voc.append(floor(center_y - (bbox_height / 2)))
            voc.append(floor(center_x + (bbox_width / 2)))
            voc.append(floor(center_y + (bbox_height / 2)))
            voc_labels.append(voc)
        create_file(file_prefix, w, h, voc_labels)
    print("Processing complete for file: {}".format(file_path))


def start():
    if not os.path.exists(DESTINATION_DIR):
        os.makedirs(DESTINATION_DIR)
    for filename in os.listdir(ANNOTATIONS_DIR_PREFIX):
        if filename.endswith('txt'):
            try:
                PathFileName = "{}/{}".format(ANNOTATIONS_DIR_PREFIX, filename)
                if os.stat(PathFileName).st_size > 0:
                    print("Si")
                    read_file(filename) 
            except:
                print("No")         
            
        else:
            print("Skipping file: {}".format(filename))


if __name__ == "__main__":
    start()

QuandoSJ · 2020-04-13T08:16:00Z

for VOC format, do we not need a xml file for a image that has no object at all? cause from the code I assume that it passed cases where txt is empty which means there is no object in the picture.

Usmansagemode · 2020-04-13T08:25:30Z

Yes Even if the class/object you want is not present in the image. You need XML which has the annotation for no object. Regards, Usman Khalid Mian, 20100061.

…

On Mon, Apr 13, 2020, 13:16 QuandoSJ ***@***.***> wrote: ***@***.**** commented on this gist. ------------------------------ for VOC format, do we not need a xml file for a image that has no object at all? cause from the code I assume that it passed cases where txt is empty which means there is no object in the picture. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <https://gist.github.com/7ca514458d24af980669b8b1c8bcdafd#gistcomment-3252146>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AL2CTU2BGVS56K6DRVDF6ODRMLC45ANCNFSM4IRCUUAQ> .

QuandoSJ · 2020-04-13T08:46:21Z

Yes Even if the class/object you want is not present in the image. You need XML which has the annotation for no object. Regards, Usman Khalid Mian, 20100061.
…
Thank you！ But the code in this repo does not generate xml for non-object image. where it ignored YOLO empty txt file, which represents the non-object images. Is there any way to fix this issue?

Usmansagemode · 2020-04-13T09:04:12Z

You need to create annotation for that. Don't create bounding box. Just annotate it as no object. And it will save that as an XML

…

On Mon, Apr 13, 2020, 13:46 QuandoSJ ***@***.***> wrote: ***@***.**** commented on this gist. ------------------------------ Yes Even if the class/object you want is not present in the image. You need XML which has the annotation for no object. Regards, Usman Khalid Mian, 20100061. … <#m_5070020001616353865_> Thank you！ But the code in this repo does not generate xml for non-object image. where it ignored YOLO empty txt file, which represents the non-object images. Is there any way to fix this issue? — You are receiving this because you commented. Reply to this email directly, view it on GitHub <https://gist.github.com/7ca514458d24af980669b8b1c8bcdafd#gistcomment-3252178>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AL2CTUY4RI4EPWIRECWGY6LRMLGOXANCNFSM4IRCUUAQ> .

QuandoSJ · 2020-04-14T07:45:29Z

You need to create annotation for that. Don't create bounding box. Just annotate it as no object. And it will save that as an XML
…

Thank you! I guess I did not phrase my question well, I have already created a yolo label for images without object, and the label file is an empty file, but the code here will omit cases when the txt file is empty, so what should I save in label.txt for yolo label when there is no object so that this code will work and transfer to an empty xml file?

gunaytemur · 2020-04-19T14:51:10Z

You need to create annotation for that. Don't create bounding box. Just annotate it as no object. And it will save that as an XML
…

Thank you! I guess I did not phrase my question well, I have already created a yolo label for images without object, and the label file is an empty file, but the code here will omit cases when the txt file is empty, so what should I save in label.txt for yolo label when there is no object so that this code will work and transfer to an empty xml file?

A useful tip:

Make sure that you have all your images in a folder name images in the same directory where you execute this script

I would like to suggest one edit on line #78
Change
with open(file_path, 'r') as file:
To
with open(os.path.join(ANNOTATIONS_DIR_PREFIX,file_path), 'r') as file:

Otherwise script generates an error, as it tries to read the filename (label .txt file) without complete path.

Absolutely right. Congratulations. Code is working with this plugin. Also, it is mandatory to use a round in voc.append calculation. XML does not accept comma numbers. labeling map is not available with this reason. If;
voc.append (round (center_x - (bbox_width / 2)))
             voc.append (round (center_y - (bbox_height / 2)))
             voc.append (round (center_x + (bbox_width / 2)))
             voc.append (round (center_y + (bbox_height / 2)))

If you use the form, your data will be easily tagged.

gunaytemur · 2020-04-19T14:56:03Z

full working of code. thanks @goodhamgupta

import os
import xml.etree.cElementTree as ET
from PIL import Image

ANNOTATIONS_DIR_PREFIX = "data"

DESTINATION_DIR = "converted_labels"

CLASS_MAPPING = {
    '0': 'name'
    # Add your remaining classes here.
}


def create_root(file_prefix, width, height):
    root = ET.Element("annotation")
    ET.SubElement(root, "folder").text = "converted_labels"
    ET.SubElement(root, "filename").text = "{}.jpg".format(file_prefix)
    ET.SubElement(root, "path").text = "D:\PycharmProjects\convert\converted_labels\{}.jpg".format(file_prefix)
    source = ET.SubElement(root, "source")
    ET.SubElement(source, "database").text = "Unknown"
    size = ET.SubElement(root, "size")
    ET.SubElement(size, "width").text = str(width)
    ET.SubElement(size, "height").text = str(height)
    ET.SubElement(size, "depth").text = "3"
    ET.SubElement(root, "segmented").text = "0"
    return root


def create_object_annotation(root, voc_labels):
    for voc_label in voc_labels:
        obj = ET.SubElement(root, "object")
        ET.SubElement(obj, "name").text=str(voc_label[0])
        ET.SubElement(obj, "pose").text = "Unspecified"
        ET.SubElement(obj, "truncated").text = str(0)
        ET.SubElement(obj, "difficult").text = str(0)
        bbox = ET.SubElement(obj, "bndbox")
        ET.SubElement(bbox, "xmin").text = str(voc_label[1])
        ET.SubElement(bbox, "ymin").text = str(voc_label[2])
        ET.SubElement(bbox, "xmax").text = str(voc_label[3])
        ET.SubElement(bbox, "ymax").text = str(voc_label[4])
    return root


def create_file(file_prefix, width, height, voc_labels):
    root = create_root(file_prefix, width, height)
    root = create_object_annotation(root, voc_labels)
    tree = ET.ElementTree(root)
    tree.write("{}/{}.xml".format(DESTINATION_DIR, file_prefix))


def read_file(file_path):
    file_prefix = file_path.split(".txt")[0]
    image_file_name = "{}.jpg".format(file_prefix)
    img = Image.open("{}/{}".format("data", image_file_name))
    w, h = img.size
    with open(os.path.join(ANNOTATIONS_DIR_PREFIX,file_path), 'r') as file:
        lines = file.readlines()
        voc_labels = []
        for line in lines:
            voc = []
            line = line.strip()
            data = line.split()
            CLASS_MAPPING.get(data[0])
            a = int(data[0])
            if  a == 0:
                voc.append("al")
            else:
                voc.append("sat")
            bbox_width = float(data[3]) * w
            bbox_height = float(data[4]) * h
            center_x = float(data[1]) * w
            center_y = float(data[2]) * h
            voc.append(round(center_x - (bbox_width / 2)))
            voc.append(round(center_y - (bbox_height / 2)))
            voc.append(round(center_x + (bbox_width / 2)))
            voc.append(round(center_y + (bbox_height / 2)))
            voc_labels.append(voc)
        create_file(file_prefix, w, h, voc_labels)
    print("Processing complete for file: {}".format(file_path))


def start():
    if not os.path.exists(DESTINATION_DIR):
        os.makedirs(DESTINATION_DIR)
    for filename in os.listdir(ANNOTATIONS_DIR_PREFIX):
        if filename.endswith('txt'):
            read_file(filename)
        else:
            print("Skipping file: {}".format(filename))


if __name__ == "__main__":
    start()

adamfahmi48 · 2020-05-31T20:22:04Z

Thanks @goodhamgupta @mjahanifar

Murtaza-Kazmi · 2020-07-18T03:07:23Z

hello. Just change the Annotations_dir_prefix and DESTINATION_DIR files and your code will run perfectly given that YOU HAVE THE ANNOTATIONS AND IMAGES IN THE SAME FOLDER.

`import os
import xml.etree.cElementTree as ET
from PIL import Image

ANNOTATIONS_DIR_PREFIX = "C:/Users/Murtaza Kazmi/Folder/"

DESTINATION_DIR = "C:/Users/Murtaza Kazmi/Corrected_Annotations_to_XML/"

CLASS_MAPPING = {
'0': '0',
'1':'1',
'2': '2',
'3': '3',
'4': '4',
'5': '5',
'6':'6',
'7':'7',
'8':'8',
'9':'9',
# Add your remaining classes here.
}

def create_root(file_prefix, width, height):
root = ET.Element("annotations")
ET.SubElement(root, "filename").text = "{}.jpg".format(file_prefix)
ET.SubElement(root, "folder").text = "images"
size = ET.SubElement(root, "size")
ET.SubElement(size, "width").text = str(width)
ET.SubElement(size, "height").text = str(height)
ET.SubElement(size, "depth").text = "3"
return root

def create_object_annotation(root, voc_labels):
for voc_label in voc_labels:
obj = ET.SubElement(root, "object")
ET.SubElement(obj, "name").text = voc_label[0]
ET.SubElement(obj, "pose").text = "Unspecified"
ET.SubElement(obj, "truncated").text = str(0)
ET.SubElement(obj, "difficult").text = str(0)
bbox = ET.SubElement(obj, "bndbox")
ET.SubElement(bbox, "xmin").text = str(voc_label[1])
ET.SubElement(bbox, "ymin").text = str(voc_label[2])
ET.SubElement(bbox, "xmax").text = str(voc_label[3])
ET.SubElement(bbox, "ymax").text = str(voc_label[4])
return root

def create_file(file_prefix, width, height, voc_labels):
root = create_root(file_prefix, width, height)
root = create_object_annotation(root, voc_labels)
tree = ET.ElementTree(root)
tree.write("{}/{}.xml".format(DESTINATION_DIR, file_prefix))

def read_file(file_path):
file_prefix = file_path.split(".txt")[0]
image_file_name = "{}.jpg".format(file_prefix)
img = Image.open(image_file_name)
w, h = img.size
print(file_path, '101')
with open(ANNOTATIONS_DIR_PREFIX + file_path, 'r') as file:
lines = file.readlines()
voc_labels = []
for line in lines:
voc = []
line = line.strip()
data = line.split()
voc.append(CLASS_MAPPING.get(data[0]))
bbox_width = float(data[3]) * w
bbox_height = float(data[4]) * h
center_x = float(data[1]) * w
center_y = float(data[2]) * h
voc.append(center_x - (bbox_width / 2))
voc.append(center_y - (bbox_height / 2))
voc.append(center_x + (bbox_width / 2))
voc.append(center_y + (bbox_height / 2))
voc_labels.append(voc)
create_file(file_prefix, w, h, voc_labels)
print("Processing complete for file: {}".format(file_path))

def start():
if not os.path.exists(DESTINATION_DIR):
os.makedirs(DESTINATION_DIR)
for filename in os.listdir(ANNOTATIONS_DIR_PREFIX):
if filename.endswith('txt'):
read_file(filename)
else:
print("Skipping file: {}".format(filename))

if name == "main":
start()
`

ekesdf · 2021-01-11T15:40:11Z

this code works

import os
from xml.dom import minidom
import xml.etree.cElementTree as ET
from PIL import Image

ANNOTATIONS_DIR_PREFIX = "path to your yolo annotations"

DESTINATION_DIR = "output path"

CLASS_MAPPING = {
    '1': 'Char'
}

def formatter(elem):
    """Return a pretty-printed XML string for the Element.
    """
    rough_string = ET.tostring(elem, 'utf-8')
    reparsed = minidom.parseString(rough_string)
    return reparsed.toprettyxml(indent="    ")

def create_root(file_prefix, width, height):
    root = ET.Element("annotation")
    ET.SubElement(root, "filename").text = "{}.jpg".format(file_prefix)
    size = ET.SubElement(root, "size")
    ET.SubElement(size, "width").text = str(width)
    ET.SubElement(size, "height").text = str(height)
    ET.SubElement(size, "depth").text = "3"
    return root


def create_object_annotation(root, voc_labels):
    for voc_label in voc_labels:
        obj = ET.SubElement(root, "object")
        ET.SubElement(obj, "name").text = voc_label[0]
        bbox = ET.SubElement(obj, "bndbox")
        ET.SubElement(bbox, "xmin").text = str(voc_label[1])
        ET.SubElement(bbox, "ymin").text = str(voc_label[2])
        ET.SubElement(bbox, "xmax").text = str(voc_label[3])
        ET.SubElement(bbox, "ymax").text = str(voc_label[4])
    return root


def create_file(file_prefix, width, height, voc_labels):
    root = create_root(file_prefix, width, height)
    root = create_object_annotation(root, voc_labels)
    with open("{}/{}.xml".format(DESTINATION_DIR, file_prefix), "w") as f:

            f.write(formatter(root))
            f.close()

def read_file(file_path):
    file_prefix = file_path.split(".txt")[0]
    image_file_name = "{}.jpg".format(file_prefix)
    img = Image.open("{}/{}".format("sites", image_file_name))
    w, h = img.size
    with open("labels/"+file_path, 'r') as file:
        lines = file.readlines()
        voc_labels = []
        for line in lines:
            voc = []
            line = line.strip()
            data = line.split()
            voc.append(CLASS_MAPPING.get(data[0]))
            bbox_width = float(data[3]) * w
            bbox_height = float(data[4]) * h
            center_x = float(data[1]) * w
            center_y = float(data[2]) * h
            voc.append(round(center_x - (bbox_width / 2)))
            voc.append(round(center_y - (bbox_height / 2)))
            voc.append(round(center_x + (bbox_width / 2)))
            voc.append(round(center_y + (bbox_height / 2)))
            voc_labels.append(voc)
        create_file(file_prefix, w, h, voc_labels)


def start():
    if not os.path.exists(DESTINATION_DIR):
        os.makedirs(DESTINATION_DIR)
    for filename in os.listdir(ANNOTATIONS_DIR_PREFIX):
        if filename.endswith('txt'):
            read_file(filename)
        else:
            print("Skipping file: {}".format(filename))


if __name__ == "__main__":
    start()

goodhamgupta/yolo_to_voc.py

isennkubilay commented Sep 27, 2019

Usmansagemode commented Oct 21, 2019

AdonaiVera commented Oct 31, 2019 •

edited

Loading

devika-28 commented Nov 7, 2019

devika-28 commented Nov 7, 2019

Rutvikshah47 commented Nov 14, 2019

HiddenFox29 commented Nov 20, 2019

khalidw commented Dec 9, 2019 •

edited

Loading

SiBensberg commented Jan 21, 2020 •

edited

Loading

ashwath1295 commented Mar 24, 2020

vikashkrs commented Apr 2, 2020 •

edited

Loading

mjahanifar commented Apr 9, 2020

QuandoSJ commented Apr 13, 2020

Usmansagemode commented Apr 13, 2020 via email

QuandoSJ commented Apr 13, 2020

Usmansagemode commented Apr 13, 2020 via email

QuandoSJ commented Apr 14, 2020

gunaytemur commented Apr 19, 2020

gunaytemur commented Apr 19, 2020

adamfahmi48 commented May 31, 2020

Murtaza-Kazmi commented Jul 18, 2020 •

edited

Loading

ekesdf commented Jan 11, 2021

	# Script to convert yolo annotations to voc format

	# Sample format
	# <annotation>
	# <folder>_image_fashion</folder>
	# <filename>brooke-cagle-39574.jpg</filename>
	# <size>
	# <width>1200</width>
	# <height>800</height>
	# <depth>3</depth>
	# </size>
	# <segmented>0</segmented>
	# <object>
	# <name>head</name>
	# <pose>Unspecified</pose>
	# <truncated>0</truncated>
	# <difficult>0</difficult>
	# <bndbox>
	# <xmin>549</xmin>
	# <ymin>251</ymin>
	# <xmax>625</xmax>
	# <ymax>335</ymax>
	# </bndbox>
	# </object>
	# <annotation>
	import os
	import xml.etree.cElementTree as ET
	from PIL import Image

	ANNOTATIONS_DIR_PREFIX = "annotations"

	DESTINATION_DIR = "converted_labels"

	CLASS_MAPPING = {
	'0': 'name'
	# Add your remaining classes here.
	}


	def create_root(file_prefix, width, height):
	root = ET.Element("annotations")
	ET.SubElement(root, "filename").text = "{}.jpg".format(file_prefix)
	ET.SubElement(root, "folder").text = "images"
	size = ET.SubElement(root, "size")
	ET.SubElement(size, "width").text = str(width)
	ET.SubElement(size, "height").text = str(height)
	ET.SubElement(size, "depth").text = "3"
	return root


	def create_object_annotation(root, voc_labels):
	for voc_label in voc_labels:
	obj = ET.SubElement(root, "object")
	ET.SubElement(obj, "name").text = voc_label[0]
	ET.SubElement(obj, "pose").text = "Unspecified"
	ET.SubElement(obj, "truncated").text = str(0)
	ET.SubElement(obj, "difficult").text = str(0)
	bbox = ET.SubElement(obj, "bndbox")
	ET.SubElement(bbox, "xmin").text = str(voc_label[1])
	ET.SubElement(bbox, "ymin").text = str(voc_label[2])
	ET.SubElement(bbox, "xmax").text = str(voc_label[3])
	ET.SubElement(bbox, "ymax").text = str(voc_label[4])
	return root


	def create_file(file_prefix, width, height, voc_labels):
	root = create_root(file_prefix, width, height)
	root = create_object_annotation(root, voc_labels)
	tree = ET.ElementTree(root)
	tree.write("{}/{}.xml".format(DESTINATION_DIR, file_prefix))


	def read_file(file_path):
	file_prefix = file_path.split(".txt")[0]
	image_file_name = "{}.jpg".format(file_prefix)
	img = Image.open("{}/{}".format("images", image_file_name))
	w, h = img.size
	with open(file_path, 'r') as file:
	lines = file.readlines()
	voc_labels = []
	for line in lines:
	voc = []
	line = line.strip()
	data = line.split()
	voc.append(CLASS_MAPPING.get(data[0]))
	bbox_width = float(data[3]) * w
	bbox_height = float(data[4]) * h
	center_x = float(data[1]) * w
	center_y = float(data[2]) * h
	voc.append(center_x - (bbox_width / 2))
	voc.append(center_y - (bbox_height / 2))
	voc.append(center_x + (bbox_width / 2))
	voc.append(center_y + (bbox_height / 2))
	voc_labels.append(voc)
	create_file(file_prefix, w, h, voc_labels)
	print("Processing complete for file: {}".format(file_path))


	def start(dir_name):
	if not os.path.exists(DESTINATION_DIR):
	os.makedirs(DESTINATION_DIR)
	for filename in os.listdir(ANNOTATIONS_DIR_PREFIX):
	if filename.endswith('txt'):
	read_file(filename)
	else:
	print("Skipping file: {}".format(filename))


	if __name__ == "__main__":
	start()

goodhamgupta/yolo_to_voc.py

isennkubilay commented Sep 27, 2019

Usmansagemode commented Oct 21, 2019

AdonaiVera commented Oct 31, 2019 • edited Loading

devika-28 commented Nov 7, 2019

devika-28 commented Nov 7, 2019

Rutvikshah47 commented Nov 14, 2019

HiddenFox29 commented Nov 20, 2019

khalidw commented Dec 9, 2019 • edited Loading

SiBensberg commented Jan 21, 2020 • edited Loading

ashwath1295 commented Mar 24, 2020

vikashkrs commented Apr 2, 2020 • edited Loading

mjahanifar commented Apr 9, 2020

QuandoSJ commented Apr 13, 2020

Usmansagemode commented Apr 13, 2020 via email

QuandoSJ commented Apr 13, 2020

Usmansagemode commented Apr 13, 2020 via email

QuandoSJ commented Apr 14, 2020

gunaytemur commented Apr 19, 2020

gunaytemur commented Apr 19, 2020

adamfahmi48 commented May 31, 2020

Murtaza-Kazmi commented Jul 18, 2020 • edited Loading

ekesdf commented Jan 11, 2021

AdonaiVera commented Oct 31, 2019 •

edited

Loading

khalidw commented Dec 9, 2019 •

edited

Loading

SiBensberg commented Jan 21, 2020 •

edited

Loading

vikashkrs commented Apr 2, 2020 •

edited

Loading

Murtaza-Kazmi commented Jul 18, 2020 •

edited

Loading