match coco evaluate shape and creat own get_image_id function

README.md

@@ -4,6 +4,10 @@
 ![](https://img.shields.io/static/v1?label=pytorch&message=1.4&color=<COLOR>)
 [![](https://img.shields.io/static/v1?label=license&message=Apache2&color=green)](./License.txt)
 
+-Add own xml to yolov4 train.txt.
+
+-match the coco evaluate shape requirement.
+
 A minimal PyTorch implementation of YOLOv4.
 - Paper Yolo v4: https://arxiv.org/abs/2004.10934
 - Source code:https://github.com/AlexeyAB/darknet

dataset.py

@@ -423,6 +423,7 @@ def get_image_id(filename:str) -> int:
     >>> no = f"{int(no):04d}"
     >>> return int(lv+no)
     """
+
     # raise NotImplementedError("Create your own 'get_image_id' function")
     # lv, no = os.path.splitext(os.path.basename(filename))[0].split("_")
     # lv = lv.replace("level", "")
@@ -437,6 +438,7 @@ def get_image_id(filename:str) -> int:
     return id
 
 
+
 if __name__ == "__main__":
     from cfg import Cfg
     import matplotlib.pyplot as plt

xml2yolov4.py

@@ -0,0 +1,80 @@
+# -*- coding: utf-8 -*-
+"""
+Created on Wed Aug 26 15:05:40 2020
+
+@author: Liu qi
+"""
+
+
+from xml.dom.minidom import parse
+import xml.dom.minidom
+import os
+import shutil
+
+def get_file_path(file_path):
+    '''
+    :param filename:修改后路径
+    :param file_path: 图片所在label文件夹目录
+    :param h,w: 图片的长宽
+    :return:图片的路径列表
+    '''
+    img_paths = []
+    img_names = os.listdir(file_path)
+    for img_name in img_names:
+        img_path = os.path.join(file_path, img_name)
+        img_paths.append(img_path)
+    return img_paths
+def mkdir_if_not_exist(path):
+    if not os.path.exists(os.path.join(*path)):
+        os.makedirs(os.path.join(*path))
+    return os.path.join(*path)
+path = r'D:\bonc\DPcase\yolov4_yolov5\nonmotor\new_data_1_label'
+mkdir_if_not_exist(['./images/train/'])
+xml_paths = get_file_path(path)
+
+label_dict  = {"bicycle":0,
+               "express_bicycle":1,
+               "express_tricycle":2,
+               "tricycle":3,
+               "electric_bicycle":4,}
+f = open('train_1.txt','w+')
+
+for xml_path in xml_paths:
+    domTree = parse(xml_path)
+    rootNode = domTree.documentElement
+    #img_name
+    filename = rootNode.getElementsByTagName("filename")
+    img_name = filename[0].childNodes[0].data
+    #move image
+    #shutil.copy('D:/bonc/DPcase/non-motor vehicle/new_data_1/'+img_name,'../data/'+img_name)
+
+    rm = True
+    xmlobjects = rootNode.getElementsByTagName("object")
+    for xmlobject in xmlobjects:
+        obname = xmlobject.getElementsByTagName("name")
+        label_name = obname[0].childNodes[0].data
+        if label_name in label_dict:
+            if rm:
+                f.write('\n')
+                f.write('../nonmotor/images/train/'+img_name)
+            rm =False
+            label = label_dict[label_name]
+
+            obbox = xmlobject.getElementsByTagName("bndbox")
+
+            obxmin = obbox[0].getElementsByTagName("xmin")
+            xmin = obxmin[0].childNodes[0].data
+
+            obymin = obbox[0].getElementsByTagName("ymin")
+            ymin = obymin[0].childNodes[0].data
+
+            obxmax = obbox[0].getElementsByTagName("xmax")
+            xmax = obxmax[0].childNodes[0].data
+
+            obymax = obbox[0].getElementsByTagName("ymax")
+            ymax = obymax[0].childNodes[0].data
+
+
+            bbox_info=" %d,%d,%d,%d,%d" % (int(xmin), int(ymin), int(xmax), int(ymax), label)
+            f.write(bbox_info)
+f.close()