--- a
+++ b/dataprocess/utils.py
@@ -0,0 +1,36 @@
+import numpy as np
+import random
+import copy
+
+
+def data_resample4seg(train_datas, train_masks,index=-2, splitvalue=3, resample_rate=1, is_big=True):
+    print('Before Resample datas:', len(train_datas))
+    res_datas = copy.deepcopy(train_datas)
+    res_masks = copy.deepcopy(train_masks)
+    for train_data, train_mask in zip(train_datas,train_masks):
+        if (is_big is True and int(train_data.split('_')[index]) >= splitvalue) or (is_big is False and int(train_data.split('_')[index]) < splitvalue):
+            for i in range(resample_rate):
+                res_datas.append(train_data)
+                res_masks.append(train_mask)
+    sorted(res_datas)
+    sorted(res_masks)
+    temp = list(zip(res_datas, res_masks))
+    random.shuffle(temp)
+    res_datas, res_masks = zip(*temp)
+    print('After Resample datas:', len(res_datas))
+    return res_datas, res_masks
+
+
+def data_resample4cls(train_datas, index=-2, splitvalue=3, resample_rate=1, is_big=True):
+    print('Before Resample datas:',len(train_datas))
+    resample_datas = copy.deepcopy(train_datas)
+    for train_data in train_datas:
+        # print(train_data)
+        value = int(train_data.split('_')[index])
+        if (is_big is True and value >= splitvalue) or (is_big is False and value <= splitvalue):
+            for i in range(resample_rate):
+                resample_datas.append(train_data)
+                # print(train_data)
+    print('After Resample datas:', len(resample_datas))
+    random.shuffle(resample_datas)
+    return resample_datas