TensorFlow 对象检测：如何批量检测答案

【问题标题】：Tensorflow object detection: how to detect on batchTensorFlow 对象检测：如何批量检测
【发布时间】：2023-03-25 16:13:01
【问题描述】：

我正在尝试使用tensorflow detection tutorial 对批次执行检测，但下面的代码给了我setting an array element with a sequence. 错误。

# load multiple images 
np_images = []
for img_path in img_paths:
    img = Image.open(image_path)
    image_np = load_image_into_numpy_array(img)      
    image_np_expanded = np.expand_dims(image_np, axis=0)
    np_images.append(image_np)

#Get input and output tensors
image_tensor = det_graph.get_tensor_by_name('image_tensor:0')  
boxes = det_graph.get_tensor_by_name('detection_boxes:0')     
scores = det_graph.get_tensor_by_name('detection_scores:0')
classes = det_graph.get_tensor_by_name('detection_classes:0')
num_detections = det_graph.get_tensor_by_name('num_detections:0')

# detect on batch of images
detection_results = sess.run(
        [boxes, scores, classes, num_detections],
        feed_dict={image_tensor: np_images})

如何正确输入图像数组？

【问题讨论】：

您能否提供带有堆栈跟踪的错误消息？

标签： python tensorflow deep-learning object-detection-api

【解决方案1】：

feed_dict 中的 image_tensor 的维度应该是 [batch_size, x, y, 3]，其中 (x,y) 是每个图像的大小。如果您的图像大小都不同，则无法创建这样的 numpy 数组。您可以调整图像大小来解决此问题。

# If the NN was trained on (300,300) size images
IMAGE_SIZE = (300, 300)
for img_path in img_paths:
    img = Image.open(image_path).resize(IMAGE_SIZE)
    image_np = load_image_into_numpy_array(img)      
    np_images.append(image_np)
...
detection_results = sess.run(
    [boxes, scores, classes, num_detections],
    feed_dict={image_tensor: np.array(np_images)})

【讨论】：