134 lines
4.4 KiB
Python
Executable File
134 lines
4.4 KiB
Python
Executable File
import os
|
|
import glob
|
|
from torchvision.datasets import ImageFolder
|
|
import torchvision.transforms as transforms
|
|
from torch.utils.data import Subset, DataLoader
|
|
import numpy as np
|
|
|
|
class Load_Data_Prepare:
|
|
def __init__(self) -> None:
|
|
self.__Label_List = []
|
|
self.__Data_List = []
|
|
self.__Contect_Dictionary = {}
|
|
self.__Final_Dict_data = {}
|
|
self.__PreSave_Data_Root = [] # 所有要讀取資料所在的位置
|
|
self.__Data_Content = []
|
|
pass
|
|
|
|
def Set_Data_Content(self, Content, Length):
|
|
tmp = []
|
|
for i in range(Length):
|
|
tmp.append(Content)
|
|
|
|
self.__Data_Content = tmp
|
|
|
|
def Set_Label_List(self, Label_List): # 為讀取檔案準備label list
|
|
self.__Label_List = Label_List
|
|
pass
|
|
|
|
def Set_Data_List(self, Data_List):
|
|
self.__Data_List = Data_List
|
|
pass
|
|
|
|
def Set_Data_Dictionary(self, Label : list, Content : list, Total_Label_Size : int):
|
|
'''將資料合併成1個Dict'''
|
|
for i in range(Total_Label_Size):
|
|
temp = {Label[i] : Content[i]}
|
|
self.__Contect_Dictionary.update(temp)
|
|
pass
|
|
|
|
def Set_Final_Dict_Data(self, Name : list, Label_Root : list, Label_LabelEncoding : list, Label_Len : int):
|
|
'''
|
|
Name : 讀取出來的Data Root的名字
|
|
Label_Root: 所有影像資料的路徑
|
|
Label_LabelEncoding: LabelEncoding後的資料
|
|
Label_Len: Label的大小
|
|
'''
|
|
for i in range(Label_Len):
|
|
temp = {Name[i] + "_Data_Root" : Label_Root[Name[i]]}
|
|
self.__Final_Dict_data.update(temp)
|
|
|
|
for i in range(Label_Len):
|
|
temp = {Name[i] + "_Data_LabelEncoding" : Label_LabelEncoding[i]}
|
|
self.__Final_Dict_data.update(temp)
|
|
|
|
def Set_PreSave_Data_Root(self, PreSave_Roots : list):
|
|
for Root in PreSave_Roots:
|
|
self.__PreSave_Data_Root.append(Root)
|
|
|
|
def Get_Label_List(self):
|
|
'''
|
|
將private的資料讀取出來
|
|
現在要放入需要的Label 需要先Set Label
|
|
'''
|
|
return self.__Label_List
|
|
|
|
def Get_Data_List(self):
|
|
return self.__Data_List
|
|
|
|
def Get_Data_Dict(self):
|
|
return self.__Contect_Dictionary
|
|
|
|
def Get_Final_Data_Dict(self):
|
|
return self.__Final_Dict_data
|
|
|
|
def Get_PreSave_Data_Root(self):
|
|
return self.__PreSave_Data_Root
|
|
|
|
def Get_Data_Content(self):
|
|
return self.__Data_Content
|
|
|
|
class Load_Data_Tools():
|
|
def __init__(self) -> None:
|
|
pass
|
|
|
|
def get_data_root(self, root, data_dict, classify_label, judge = True) -> dict :
|
|
'''取得資料路徑'''
|
|
for label in classify_label:
|
|
if judge:
|
|
path = os.path.join(root, label, "*")
|
|
else:
|
|
path = os.path.join(root, "*")
|
|
path = glob.glob(path)
|
|
data_dict[label] = path
|
|
return data_dict
|
|
|
|
def Load_ImageFolder_Data(self, Loading_Root, transform = None):
|
|
# 資料預處理
|
|
if transform == None:
|
|
transform = transforms.Compose([
|
|
transforms.Resize((512, 512))
|
|
])
|
|
else:
|
|
transform = transforms.Compose([
|
|
transforms.Resize((512, 512)),
|
|
transforms.ToTensor()
|
|
])
|
|
|
|
DataSet = ImageFolder(root=Loading_Root, transform=transform)
|
|
|
|
return DataSet
|
|
|
|
def Get_Balance_Data(self, Image_Folder : ImageFolder, total_Image_List):
|
|
class_counts = np.bincount(total_Image_List)
|
|
print("欄位大小: " + str(total_Image_List))
|
|
min_class_count = class_counts.min()
|
|
|
|
# 創建每個類別的索引並下採樣
|
|
balanced_indices = []
|
|
for class_idx in range(len(class_counts)):
|
|
class_indices = np.where(np.array(total_Image_List) == class_idx)[0]
|
|
sampled_indices = np.random.choice(class_indices, min_class_count, replace=False)
|
|
balanced_indices.extend(sampled_indices)
|
|
|
|
# 創建平衡的子集
|
|
Training_Data = Subset(Image_Folder, balanced_indices)
|
|
|
|
# 輸出內容
|
|
print(f"平衡後的每類樣本數：{min_class_count}")
|
|
|
|
return Training_Data
|
|
|
|
def DataLoad_Image_Root(self, ImageLoad, Batch):
|
|
dataloader = DataLoader(dataset = ImageLoad, batch_size = Batch, shuffle=True, num_workers = 0, pin_memory=True)
|
|
return dataloader |