fastapi/backend/apps/service/image_inference.py

#!/usr/bin/env python
"""
@Contact :   liuyuqi.gov@msn.cn
@Time    :   2024/03/25 11:34:35
@License :   Copyright © 2017-2022 liuyuqi. All Rights Reserved.
@Desc    :   image interfence
"""

# import cv2
import ssl

import numpy as np
import tensorflow as tf
from PIL import Image
from tensorflow.keras.applications.imagenet_utils import decode_predictions

ssl._create_default_https_context = ssl._create_unverified_context
classifier_model = None


class ImageClassificationService:
    """ """

    def __init__(self):
        """ """
        #    self.logger = logging.getLogger(__name__)
        self.model = ImageInferenceTask()
        self.image_classification_model = "INCEPTION_V2"
        self.IMAGE_SHAPE = (224, 224)

    async def classify(self, image_file):
        label = await self.model.predict(
            classifier_model_name=self.image_classification_model,
            image=image_file,
            shape=self.IMAGE_SHAPE,
        )
        return label


class ImageInferenceTask:
    @staticmethod
    async def download_model(cls, classifier_model_name):
        pass

    @staticmethod
    async def load_model(cls, classifier_model_name):
        """ """
        if classifier_model_name == "INCEPTION_V3":
            model = tf.keras.applications.InceptionV3(weights="imagenet")
        else:
            model = tf.keras.applications.MobileNetV2(weights="imagenet")
        return model

    async def predict(self, image_path):
        """ """
        # image = cv2.imread(image_path)
        image = Image.open(image_path)
        image = np.array(image)
        image = tf.image.resize(image, (224, 224))
        image = tf.expand_dims(image, axis=0)
        image = image / 255.0
        predictions = classifier_model.predict(image)
        predictions = decode_predictions(predictions, top=1)
        return predictions