68 lines
1.9 KiB
Python
68 lines
1.9 KiB
Python
# from keras import backend as K
|
|
# from keras.layers import Layer
|
|
import numpy as np
|
|
import os.path
|
|
import cv2
|
|
|
|
|
|
# class TripletLossLayer(Layer):
|
|
# def __init__(self, alpha, **kwargs):
|
|
# self.alpha = alpha
|
|
# super(TripletLossLayer, self).__init__(**kwargs)
|
|
#
|
|
# def triplet_loss(self, inputs):
|
|
# a, p, n = inputs
|
|
# p_dist = K.sum(K.square(a - p), axis=-1)
|
|
# n_dist = K.sum(K.square(a - n), axis=-1)
|
|
# return K.sum(K.maximum(p_dist - n_dist + self.alpha, 0), axis=0)
|
|
#
|
|
# def call(self, inputs):
|
|
# loss = self.triplet_loss(inputs)
|
|
# self.add_loss(loss)
|
|
# return loss
|
|
|
|
|
|
class IdentityMetadata():
|
|
def __init__(self, base, type, name, file):
|
|
# dataset base directory
|
|
self.base = base
|
|
# identity people type
|
|
self.type = type
|
|
# identity name
|
|
self.name = name
|
|
# image file name
|
|
self.file = file
|
|
|
|
def __repr__(self):
|
|
return self.image_path()
|
|
|
|
def image_path(self):
|
|
return os.path.join(self.base, self.type, self.name, self.file)
|
|
|
|
|
|
def load_metadata(path):
|
|
metadata = []
|
|
for i in sorted(os.listdir(path)):
|
|
type = []
|
|
for j in sorted(os.listdir(os.path.join(path, i))):
|
|
person = []
|
|
for k in sorted(os.listdir(os.path.join(path + i, j))):
|
|
# Check file extension. Allow only jpg/jpeg' files.
|
|
ext = os.path.splitext(k)[1]
|
|
if ext == '.jpg' or ext == '.jpeg':
|
|
person.append(IdentityMetadata(path, i, j, k))
|
|
type.append(person)
|
|
metadata.append(type)
|
|
return np.array(metadata, dtype=object)
|
|
|
|
|
|
def load_image(path):
|
|
img = cv2.imread(path)
|
|
# OpenCV loads images with color channels
|
|
# in BGR order. So we need to reverse them
|
|
return img[..., ::-1]
|
|
|
|
|
|
def distance(emb1, emb2):
|
|
return np.sum(np.square(emb1 - emb2))
|