calories = [656, 658, 768, 836]
# 入力画像をNumpyに変換 --- (※2)
X = []
files = []
for fname in sys.argv[1:]:
img = Image.open(fname)
img = img.convert("RGB")
img = img.resize((image_size, image_size))
in_data = np.asarray(img)
X.append(in_data)
files.append(fname)
X = np.array(X)
# CNNのモデルを構築 --- (※3)
model = gyudon.build_model(X.shape[1:])
model.load_weights("./image/gyudon-model.hdf5")
# データを予測 --- (※4)
html = ""
pre = model.predict(X)
for i, p in enumerate(pre):
y = p.argmax()
print("+ 入力:", files[i])
print("| 牛丼名:", categories[y])
print("| カロリー:", calories[y])
html += """
<h3>入力:{0}</h3>
<div>
<p><img src="{1}" width=300></p>
<p>牛丼名:{2}</p>
calories = [656, 658, 768, 836]
# 입력 이미지를 Numpy로 변환하기 --- (※2)
X = []
files = []
for fname in sys.argv[1:]:
img = Image.open(fname)
img = img.convert("RGB")
img = img.resize((image_size, image_size))
in_data = np.asarray(img)
X.append(in_data)
files.append(fname)
X = np.array(X)
# CNN 모델 구축하기 --- (※3)
model = gyudon.build_model(X.shape[1:])
model.load_weights("./image/gyudon-model.hdf5")
# 데이터 예측하기 --- (※4)
html = ""
pre = model.predict(X)
for i, p in enumerate(pre):
y = p.argmax()
print("+입력:", files[i])
print("|규동 이름:", categories[y])
print("|칼로리:", calories[y])
html += """
<h3>입력:{0}</h3>
<div>
<p><img src="{1}" width=300></p>
<p>규동 이름:{2}</p>
<p>칼로리 :{3}kcal</p>