
# 필요한 라이브러리를 가져옵니다.
import tensorflow as tf
from tensorflow import keras
from tensorflow.keras.preprocessing import image
import numpy as np
from tensorflow.keras.applications.resnet50 import preprocess_input, decode_predictions
from keras.applications.imagenet_utils import preprocess_input


# 이미지를 로드하고 전처리하는 함수
def preprocess_image(image_path):
    img = image.load_img(image_path, target_size=(224, 224))
    img = image.img_to_array(img)
    img = np.expand_dims(img, axis=0)
    img = preprocess_input(img)
    return img


# 이미지 객체 인식을 수행하는 함수
def predict_objects(image_path):
    # 사전에 훈련된 ResNet50 모델을 불러옵니다.
    model = keras.applications.resnet50.ResNet50(weights='imagenet')

    # 이미지를 전처리합니다.
    img = preprocess_image(image_path)

    # 이미지를 모델에 입력하여 예측을 수행합니다.
    preds = model.predict(img)
    # 예측 결과를 해석합니다.
    decoded_preds = decode_predictions(preds, top=3)[0]
    
    # 예측 결과를 출력합니다.
    for _, label, confidence in decoded_preds:
        print(f"{label}: {confidence * 100}%")


from IPython.display import Image


# 이미지 파일 경로를 지정합니다.
image_path = 'Dog_rawPixel01.jpg'

# 이미지 표시
Image(image_path)


# 이미지 객체 인식을 수행합니다.
predict_objects(image_path)

1/1 [==============================] - 2s 2s/step
standard_poodle: 93.59152317047119%
miniature_poodle: 4.353092610836029%
golden_retriever: 0.6918197963386774%


# 이미지 파일 경로를 지정합니다.
image_path = 'life_unsplash.jpeg'

# 이미지 표시
Image(image_path)


# 이미지 객체 인식을 수행합니다.
predict_objects(image_path)

WARNING:tensorflow:5 out of the last 5 calls to <function Model.make_predict_function.<locals>.predict_function at 0x000002AA0377C550> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.
1/1 [==============================] - 2s 2s/step
mosquito_net: 26.91299319267273%
quilt: 19.8626309633255%
crib: 19.388511776924133%


# 이미지 파일 경로를 지정합니다.
image_path = 'Deer_un.jpeg'

# 이미지 표시
Image(image_path)


# 이미지 객체 인식을 수행합니다.
predict_objects(image_path)

WARNING:tensorflow:6 out of the last 6 calls to <function Model.make_predict_function.<locals>.predict_function at 0x000002AA005BE550> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has reduce_retracing=True option that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/guide/function#controlling_retracing and https://www.tensorflow.org/api_docs/python/tf/function for  more details.
1/1 [==============================] - 2s 2s/step
ibex: 55.18874526023865%
impala: 14.141763746738434%
gazelle: 11.144579201936722%

이미지 인식 샘플 코드¶

학습 내용¶