def AnalyseInputImage(queryDir, maxNo, model_name, img_canvas, root):
h5f = h5py.File(model_name,'r')
feats = h5f['dataset_1'][:]
imgNames_utf = h5f['dataset_2'][:]
imgNames = []
for i in imgNames_utf:
imgNames.append(i.decode('utf-8'))
h5f.close()
model = VGGNet()
queryVec = model.extract_feat(queryDir)
scores = np.dot(queryVec, feats.T)
rank_ID = np.argsort(scores)[::-1]
rank_score = scores[rank_ID]
maxres = int(maxNo)
imlist = [imgNames[index] for i,index in enumerate(rank_ID[0:maxres])]
file_io = FileIO()
file_io.save_obj(imlist,"imlist")
img_canvas.delete('all')
vsbar = Scrollbar(frame_canvas, orient=VERTICAL, command=img_canvas.yview)
vsbar.grid(row=0, column=1, sticky=NS)
vsbar.config(command=img_canvas.yview)
img_canvas.configure(yscrollcommand=vsbar.set)
frame_images = Frame(img_canvas, bg="grey")
img_canvas.create_window((0,0), window=frame_images, anchor='nw')
img_no = 0
max_in_row = 0
height_total = 0
for i in imlist:
basewidth = 300
img = Image.open(i)
wpercent = (basewidth/float(img.size[0]))
hsize = int((float(img.size[1])*float(wpercent)))
max_in_row = max(max_in_row, hsize)
img = img.resize((basewidth,hsize), Image.ANTIALIAS)
render = ImageTk.PhotoImage(img)
img_show = Label(frame_images, image=render, name=str(img_no))
img_show.bind("<Button-1>", ShowOriginalImage)
img_show.image = render
img_show.grid(row=img_no//3, column=img_no%3)
img_no += 1
if img_no%3==0:
height_total += max_in_row
max_in_row = 0
frame_canvas.config(height=height_total)
root.update()
img_canvas.config(scrollregion=img_canvas.bbox("all"))
def GenerateFeatureDatabase(fd_entry, ft_label, pgbar, pglabel, root):
ft_name = "feature"
fd_name = fd_entry.get()
root.update()
if os.path.exists(fd_name) and os.path.isdir(fd_name):
pass
else:
messagebox.showerror(title="文件夹读取错误", message="无法读取文件或文件夹不存在")
return
file_handler = FileHandler()
# img_list = file_handler.get_imlist(fd_name)
img_list = [os.path.join(dp, f) for dp, dn, filenames in os.walk(fd_name) for f in filenames if os.path.splitext(f)[1] == '.jpg' or os.path.splitext(f)[1] == '.jpeg']
# print(img_list)
print("--------------------------------------------------")
print(" feature extraction starts")
print("--------------------------------------------------")
feats = []
names = []
model = VGGNet()
ite_no = 0
ite_to = len(img_list)
error_occured = False
if getattr(sys, 'frozen', False):
application_path = os.path.dirname(sys.executable)
elif __file__:
application_path = os.path.dirname(__file__)
for i, img_path in enumerate(img_list):
try:
ite_no += 1
norm_feat = model.extract_feat(img_path)
img_name = img_path
feats.append(norm_feat)
names.append(img_name.encode('utf-8'))
print("extracting feature from image No. %d , %d images in total" %((i+1), len(img_list)))
pgbar['value'] = (100 * ite_no) / ite_to
print(pgbar['value'])
label_content="{0:0.2f}%".format(pgbar['value'])
pglabel.configure(text=label_content)
root.update()
except Exception as e:
error_occured = True
logger_fname = os.path.join(application_path, 'error.log')
now = datetime.now()
dt_string = now.strftime("%d/%m/%Y %H:%M:%S")
with open(logger_fname, "a+", encoding='utf-8') as f:
f.write('[Error] ' + dt_string + "\n")
f.write('[Trace Back] ' + str(e) + "\n")
f.write('[Detail] ' + img_path + "\n")
feats = np.array(feats)
print("--------------------------------------------------")
print(" writing feature extraction results ...")
print("--------------------------------------------------")
if error_occured:
messagebox.showerror(title="提取特征时发生错误", message="提取特征时发生错误，程序跳过了错误文件完成特征提取，请将error.log文件，以及后台截图发给开发者。")
output = ft_name + "_" + hashlib.md5(fd_name.encode()).hexdigest() + ".h5"
output = os.path.join(application_path, output)
h5f = h5py.File(output, 'w')
h5f.create_dataset('dataset_1', data = feats)
h5f.create_dataset('dataset_2', data = np.string_(names))
h5f.close()
ft_label.configure(text=os.path.abspath(output))
# Load array from file and update the file path array
file_io = FileIO()
if file_io.check_obj("save"):
pic_dataset_paths = file_io.load_obj("save")
ft_existed = 0
for path in pic_dataset_paths:
if path == fd_name:
break
ft_existed += 1
if ft_existed >= len(pic_dataset_paths):
pic_dataset_paths.append(fd_name)
else:
tmp = pic_dataset_paths[ft_existed]
pic_dataset_paths[ft_existed] = pic_dataset_paths[-1]
pic_dataset_paths[-1] = tmp
print(pic_dataset_paths)
else:
pic_dataset_paths = [fd_name]
file_io.save_obj(pic_dataset_paths,"save")
fd_entry["values"] = pic_dataset_paths
fd_entry.current(len(pic_dataset_paths) - 1)