是的，AWS Textract可以捕获PDF文本的特定部分。以下是使用AWS Textract和Python示例代码：

import boto3

# AWS Textract client
textract = boto3.client('textract')

# S3 bucket name and file name
s3_bucket_name = 'my-s3-bucket'
document_name = 'sample.pdf'

# Define the bounding box of the specific portion of the document you want to extract
# Coordinates should be in the format of {'X': int, 'Y': int}. All coordinates are in document pixels (72 DPI)
bounding_box = {
    'Top': 400,
    'Left': 100,
    'Width': 300,
    'Height': 100
}

# Set up the analysis features
feature_types = ['TABLES', 'FORMS']

# Call AWS Textract to analyze the document
response = textract.analyze_document(
    Document={
        'S3Object': {
            'Bucket': s3_bucket_name,
            'Name': document_name,
        }
    },
    FeatureTypes=feature_types,
    BlockType='LINE',
    Geometry={
        'BoundingBox': bounding_box
    }
)

# Extract the text from the specific region
text = ''
for block in response['Blocks']:
    if block['BlockType'] == 'LINE':
        text += block['Text'] + '\n'

print(text)

在此示例中，我们使用了一个名为“sample.pdf”的文件并且按照我们所需的尺寸和位置定义了特定部分。运行脚本后，从PDF文档的指定位置提取文本并将其打印到控制台。