項目地址：poly-yolo
論文地址：poly-yolo論文

1、 Format of data for training

Generally, YOLO uses notation of one image per line. One line includes all the boxes inside an image.

path_to\image1.jpg x1,y1,x2,y2,class,p1x,p1y,pnx,pny x1,y1,x2,y2,class,p1x,p1y,pnx,pny path_to\image2.jpg x1,y1,x2,y2,class,p1x,p1y,pnx,pny

Where x1,y1 denote top-left of a bounding box and x2,y2 denote bottom-right. p1x,p1y … pnx,pny are coordinates of bounding box vertices.
Script labels_to_yolo_format.py converts IDD and Cityscapes dataset annotations to yolo format. The generated annotation file is put to the provided image folder. Use ‘–help’ for script parameters description.

2、 訓練網絡結構

2.1 首先需要準備數據集。

我們將coco的 val2014數據集轉成poly-yolo需要的數據集，腳本如下。將val2014的coco數據集通過下面數據集轉換之后就可以得到一個train.txt存儲這我們需要的訓練標簽。標簽格式（path_to\image1.jpg x1,y1,x2,y2,class,p1x,p1y,pnx,pny x1,y1,x2,y2,class,p1x,p1y,pnx,pny）

import json from collections import defaultdictname_box_id = defaultdict(list) name_segmentation_id = defaultdict(list) id_name = dict() f = open("instances_val2014.json",encoding='utf-8') data = json.load(f)annotations = data['annotations'] for ant in annotations:id = ant['image_id']name = 'coco/train2014/COCO_val2014_%012d.jpg' % idcat = ant['category_id']if cat >= 1 and cat <= 11:cat = cat - 1elif cat >= 13 and cat <= 25:cat = cat - 2elif cat >= 27 and cat <= 28:cat = cat - 3elif cat >= 31 and cat <= 44:cat = cat - 5elif cat >= 46 and cat <= 65:cat = cat - 6elif cat == 67:cat = cat - 7elif cat == 70:cat = cat - 9elif cat >= 72 and cat <= 82:cat = cat - 10elif cat >= 84 and cat <= 90:cat = cat - 11name_box_id[name].append([ant['bbox'], cat,ant['segmentation']])f = open('train.txt', 'w') for key in name_box_id.keys():f.write(key)box_infos = name_box_id[key]for info in box_infos:x_min = int(info[0][0])y_min = int(info[0][1])x_max = x_min + int(info[0][2])y_max = y_min + int(info[0][3])box_info = " %d,%d,%d,%d,%d," % (x_min, y_min, x_max, y_max, int(info[1]))#print(info[2])#print('*********************************')if isinstance(info[2],list):if len(info[2])==1:f.write(box_info)lista = []for i in info[2][0]:i = int(i)lista.append(i)f.write(str(lista)) f.write('\n') f.close()

2.2將calss類別修改成coco的80類別，運行訓練模型。

python poly-yolo.py

網絡就開始訓練了。

參考：將POLY-YOLO代碼跑起來的環境配置，poly-yolo訓練自己的數據集

總結

以上是生活随笔為你收集整理的poly-yolo训练自己的数据的全部內容，希望文章能夠幫你解決所遇到的問題。

如果覺得生活随笔網站內容還不錯，歡迎將生活随笔推薦給好友。