Question

我正在尝试将图像转换为分类，但是找不到灵魂。

实际代码将转换一张图片，而不是整个文件夹。是否有任何想从文件夹“ images”中读取所有图像并将其转换为具有骨骼结构的文件夹“ output”的想法？

非常感谢。

亚历山大

import logging
import time
import os
import cv2
import numpy as np

from tf_pose.estimator import TfPoseEstimator
from tf_pose.networks import get_graph_path, model_wh
import scripts.label_image as label_img
import scripts.label_image_scene as label_img_scene

logger = logging.getLogger('Pose_Action_and_Scene_Understanding')
logger.setLevel(logging.DEBUG)
ch = logging.StreamHandler()
ch.setLevel(logging.DEBUG)
formatter = logging.Formatter('[%(asctime)s] [%(name)s] [%(levelname)s] %(message)s')
ch.setFormatter(formatter)
logger.addHandler(ch)

fps_time = 0
address = os.getcwd()
if __name__ == '__main__':
    parser = argparse.ArgumentParser(description='tf-classification')
    parser.add_argument('--image', type=str, required=True)
    parser.add_argument('--show-process', type=bool, default=False,
                        help='for debug purpose, if enabled, speed for inference is dropped.')
    args = parser.parse_args()

    logger.debug('initialization %s : %s' % ('mobilenet_thin', get_graph_path('mobilenet_thin')))
    e = TfPoseEstimator(get_graph_path('mobilenet_thin'), target_size=(432, 368))
    image = cv2.imread(args.image)
    logger.info('cam image=%dx%d' % (image.shape[1], image.shape[0]))
    
    logger.debug('+image processing+')
    logger.debug('+postprocessing+')
    start_time = time.time()
    humans = e.inference(image, upsample_size=4.0)
    img = TfPoseEstimator.draw_humans(image, humans, imgcopy=False)
    
    logger.debug('+classification+')
    # Getting only the skeletal structure (with white background) of the actual image
    image = np.zeros(image.shape,dtype=np.uint8)
    image.fill(255) 
    image = TfPoseEstimator.draw_humans(image, humans, imgcopy=False)
    
    cv2.imwrite('show1.png',image)
    cv2.imshow('tf-classification', image)
    cv2.waitKey(0)
    logger.debug('+finished+')
    cv2.destroyAllWindows()

Answer 1

根据数据的组织方式，可以使用ImageDataGenerator扩展图像。然后使用.flow或.flow_from_directory将转换后的图像保存到您指定的目录中。文档为here.

用于分类的图像转换器

1 个答案: