106062515劉臨軒

evaluate_acc.py

@@ -0,0 +1,28 @@
+# Evalute the accuracy
+
+import numpy as np
+from load_datas import *
+import pdb
+
+
+# ground truth label path
+GT_label_path = '../../labels/'
+
+# prediction label path
+pred_label_path = '../../labels/'
+
+#load_train_labels(label_path,'obj',setting_index=0)
+GT  =load_test_labels(GT_label_path, 'obj',setting_index=0)
+pred=load_test_labels(pred_label_path, 'obj',setting_index=0)
+
+
+GT_seq=np.zeros((0))
+pred_seq=np.zeros((0))
+
+for i in xrange(len(GT)):
+    GT_seq=np.append(GT_seq, GT[i])
+    pred_seq  =np.append(pred_seq, pred[i])
+
+correct = np.count_nonzero(GT_seq==pred_seq)
+total = len(GT_seq)
+print "testing acc. = %.2f%%, %d/%d" %(correct/float(total)*100, correct, total)

load_datas.py

@@ -0,0 +1,102 @@
+#############################################################################################################
+# load features and labels                                                                                  #
+# feature dimension : (N,d), (i.e., N*4096), label dimension : (N,)                                         #
+# type of all the output will be "list"                                                                     #
+#############################################################################################################
+# There are two different setting of training and testing :                                                 #
+# The default setting (setting-0) is splitting the dataset into two parts, half for training, half for      #
+# testing. Training index: lab-1234, office-123, house-123. Testing index: lab-5678, office-456, house-456  #
+#############################################################################################################
+
+import sys, os
+import numpy as np
+import pdb
+import warnings
+
+def load_one_label_seq(path):
+    npy = np.load(path)
+    #print path
+    return npy
+
+
+def load_label_seqs(path, mode, index):
+    labels=[]
+    for i in range(len(index)):
+        loc = index[i][0]
+        idx = index[i][1]
+        labelnpy = os.path.join(path,loc,mode+"_left"+str(idx)+'.npy')
+        labels.append(load_one_label_seq(labelnpy))
+        labelnpy = os.path.join(path,loc,mode+"_right"+str(idx)+'.npy')
+        labels.append(load_one_label_seq(labelnpy))
+    return labels
+
+def gen_index(setting_index):
+    train_index=[]
+    test_index =[]
+    if setting_index == 0:
+        for i in range(1,9):
+            if i <= 4:
+                train_index.append(('lab',i))
+            else:
+                test_index.append(('lab',i))
+        for i in range(1,7):
+            if i <= 3:
+                train_index.append(('office',i))
+            else:
+                test_index.append(('office',i))
+        for i in range(1,7):
+            if i <= 3:
+                train_index.append(('house',i))
+            else:
+                test_index.append(('house',i))
+    elif setting_index == 1:
+        for i in range(1,9):
+            train_index.append(('lab',i))
+        for i in range(1,7):
+            train_index.append(('office',i))
+        for i in range(1,7):
+            test_index.append(('house',i))
+    else:
+        raise ValueError ('error setting index')
+
+    return train_index, test_index
+
+
+
+def gen_index_process(index=None, setting_index=None):
+    if index == None:
+        if setting_index==None:
+            raise ValueError('Setting index can not be none')
+        else:
+            train_index, test_index = gen_index(setting_index)
+    return train_index, test_index
+
+
+def load_train_labels(path, mode, index=None, setting_index=None):
+    if index == None:
+        index,_ = gen_index_process(index,setting_index)
+    else:
+        if setting_index != None:
+            warnings.warn('setting_index has no effect when given particular index')
+    labels = load_label_seqs(path, mode, index)
+    return labels
+
+def load_test_labels(path, mode, index=None, setting_index=None):
+    if index == None:
+        _,index = gen_index_process(index,setting_index)
+    else:
+        if setting_index != None:
+            warnings.warn('setting_index has no effect when given particular index')
+    labels = load_label_seqs(path, mode, index)
+    return labels
+
+
+def load_all_labels(path, mode, setting_index):
+    train_index, test_index = gen_index(setting_index)
+    train_labels = load_train_labels(path, mode,train_index)
+    test_labels = load_train_labels(path, mode,test_index)
+    return train_labels, test_labels
+
+
+
+

results/index.md

@@ -1,47 +1,19 @@
-# Your Name <span style="color:red">(id)</span>
+# Caroline <span style="color:red">(106062515)</span>
 
 #Project 5: Deep Classification
 
 ## Overview
-The project is related to 
-> quote
-
+The project is VGG model training on given images and labels.
 
 ## Implementation
-1. One
-	* item
-	* item
-2. Two
-
-```
-Code highlights
-```
+1. test_load_datas.py, load_datas.py: loads labels and training images
+2. ty.py: VGG model
 
 ## Installation
-* Other required packages.
-* How to compile from source?
+* pre-trained VGG: https://drive.google.com/open?id=0B1d4hs5v8ZQocDNXS1Njb2tQOVk
+* run python tf.py
 
 ### Results
-
-<table border=1>
-<tr>
-<td>
-<img src="placeholder.jpg" width="24%"/>
-<img src="placeholder.jpg"  width="24%"/>
-<img src="placeholder.jpg" width="24%"/>
-<img src="placeholder.jpg" width="24%"/>
-</td>
-</tr>
-
-<tr>
-<td>
-<img src="placeholder.jpg" width="24%"/>
-<img src="placeholder.jpg"  width="24%"/>
-<img src="placeholder.jpg" width="24%"/>
-<img src="placeholder.jpg" width="24%"/>
-</td>
-</tr>
-
-</table>
+After training 2 epochs, average accuracy: 0.3~0.4
 
 

test_load_datas.py

@@ -0,0 +1,139 @@
+import numpy as np
+from load_datas import *
+import pdb
+import re
+path_data = os.environ.get("GRAPE_DATASET_DIR")
+
+#label_path='/home/yoooosing/CEDL/hw1/labels/'
+label_path = os.path.join(path_data,'labels')
+
+
+# load all labels
+train_labels0, test_labels0 = load_all_labels(label_path, 'obj',0)
+#print len(train_labels0)
+#print len(test_labels0)
+# load train/test labels automatically
+#train_labels = load_train_labels(label_path,'obj',setting_index=1)
+#test_labels = load_test_labels(label_path,'obj',setting_index=1)
+
+# load particular index
+# You can load only part of the directories, the index of directories should be a list.
+# Such like [(location1, id), (location2, id) ...], ex. [('lab',1), ('lab',2), ('office',1)]
+# No setting_index required now.
+# train_labels = load_train_labels(label_path,'FA',index=[['lab',1],['lab',2],['office',3]])
+# test_labels = load_test_labels(label_path,'FA',index=[['lab',1],['lab',2],['office',3]])
+
+
+
+# load train/test images
+
+train_images = []
+duplicate_train_labels = []
+#image_path1='/home/yoooosing/CEDL/hw1/train/lab'
+#image_path2='/home/yoooosing/CEDL/hw1/train/office'
+#image_path3='/home/yoooosing/CEDL/hw1/train/house'
+
+image_path1 = os.path.join(path_data, 'frames/train/lab')
+image_path2 = os.path.join(path_data, 'frames/train/office')
+image_path3 = os.path.join(path_data, 'frames/train/house')
+
+
+
+for i in xrange(1,5):
+    temp_path_l = os.listdir(os.path.join(image_path1,str(i),'Lhand'))
+    temp_path_l = sorted(temp_path_l, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_l = [os.path.join(image_path1,str(i),'Lhand',t) for t in temp_path_l]
+    train_images.extend(temp_path_l)
+    temp_path_r = os.listdir(os.path.join(image_path1,str(i),'Rhand'))
+    temp_path_r = sorted(temp_path_r, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_r = [os.path.join(image_path1,str(i),'Rhand',t) for t in temp_path_r]
+    train_images.extend(temp_path_r)
+
+for i in xrange(1,4):
+    temp_path_l = os.listdir(os.path.join(image_path2,str(i),'Lhand'))
+    temp_path_l = sorted(temp_path_l, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_l = [os.path.join(image_path2,str(i),'Lhand',t) for t in temp_path_l]
+    train_images.extend(temp_path_l)
+    temp_path_r = os.listdir(os.path.join(image_path2,str(i),'Rhand'))
+    temp_path_r = sorted(temp_path_r, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_r = [os.path.join(image_path2,str(i),'Rhand',t) for t in temp_path_r]
+    train_images.extend(temp_path_r)
+
+for i in xrange(1,4):
+    temp_path_l = os.listdir(os.path.join(image_path3,str(i),'Lhand'))
+    temp_path_l = sorted(temp_path_l, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_l = [os.path.join(image_path3,str(i),'Lhand',t) for t in temp_path_l]
+    train_images.extend(temp_path_l)
+    temp_path_r = os.listdir(os.path.join(image_path3,str(i),'Rhand'))
+    temp_path_r = sorted(temp_path_r, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_r = [os.path.join(image_path3,str(i),'Rhand',t) for t in temp_path_r]
+    train_images.extend(temp_path_r)
+
+for i in xrange(0,8):
+    duplicate_train_labels.extend(train_labels0[i])
+for i in xrange(8,14):
+    duplicate_train_labels.extend(train_labels0[i])
+for i in xrange(14,20):
+    duplicate_train_labels.extend(train_labels0[i])
+
+test_images = []
+duplicate_test_labels = []
+#image_path1='/home/yoooosing/CEDL/hw1/test/lab'
+#image_path2='/home/yoooosing/CEDL/hw1/test/office'
+#image_path3='/home/yoooosing/CEDL/hw1/test/house'
+
+image_path1 = os.path.join(path_data, 'frames/test/lab')
+image_path2 = os.path.join(path_data, 'frames/test/office')
+image_path3 = os.path.join(path_data, 'frames/test/house')
+
+for i in xrange(1,5):
+    temp_path_l = os.listdir(os.path.join(image_path1,str(i),'Lhand'))
+    temp_path_l = sorted(temp_path_l, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_l = [os.path.join(image_path1,str(i),'Lhand',t) for t in temp_path_l]
+    test_images.extend(temp_path_l)
+    temp_path_r = os.listdir(os.path.join(image_path1,str(i),'Rhand'))
+    temp_path_r = sorted(temp_path_r, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_r = [os.path.join(image_path1,str(i),'Rhand',t) for t in temp_path_r]
+    test_images.extend(temp_path_r)
+
+for i in xrange(1,4):
+    temp_path_l = os.listdir(os.path.join(image_path2,str(i),'Lhand'))
+    temp_path_l = sorted(temp_path_l, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_l = [os.path.join(image_path2,str(i),'Lhand',t) for t in temp_path_l]
+    test_images.extend(temp_path_l)
+    temp_path_r = os.listdir(os.path.join(image_path2,str(i),'Rhand'))
+    temp_path_r = sorted(temp_path_r, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_r = [os.path.join(image_path2,str(i),'Rhand',t) for t in temp_path_r]
+    test_images.extend(temp_path_r)
+
+for i in xrange(1,4):
+    temp_path_l = os.listdir(os.path.join(image_path3,str(i),'Lhand'))
+    temp_path_l = sorted(temp_path_l, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_l = [os.path.join(image_path3,str(i),'Lhand',t) for t in temp_path_l]
+    test_images.extend(temp_path_l)
+    temp_path_r = os.listdir(os.path.join(image_path3,str(i),'Rhand'))
+    temp_path_r = sorted(temp_path_r, key=lambda x: int(re.sub('\D', '', x)))
+    temp_path_r = [os.path.join(image_path3,str(i),'Rhand',t) for t in temp_path_r]
+    test_images.extend(temp_path_r)
+
+for i in xrange(0,8):
+    duplicate_test_labels.extend(test_labels0[i])
+for i in xrange(8,14):
+    duplicate_test_labels.extend(test_labels0[i])
+for i in xrange(14,20):
+    duplicate_test_labels.extend(test_labels0[i])
+
+#print '\n'.join(train_images)
+#print len(train_images)
+#print len(duplicate_train_labels)
+#print len(test_images)
+#print len(duplicate_test_labels)
+
+
+#for i in xrange(len(train_labels)):
+#    if np.count_nonzero(train_labels0[i] == train_labels[i]) != len(train_labels0[i]):
+#        print 'error'
+
+#for i in xrange(len(test_labels)):
+#    if np.count_nonzero(test_labels0[i] == test_labels[i]) != len(test_labels[i]):
+#        print 'error'