如何查询CT机-DICOM
我正在为医院开发 dicom 图像存档和查看器。两周以来我们已经学习并了解了它们是什么！ 。仍然有一个如何查询模式（语法）的问题。我有标签表，但听到了 SCP 和 SCU 的匹配密钥和响应密钥之类的内容。
I am developing a dicom image archiving and viewer for hospitals. Since 2 weeks have studied and learned what are they ! . Still have a question of how to query the modalities (Syntax). I have the tag sheet but heard something like matching key and response key from SCP and SCU.
如果你对这篇内容有疑问，欢迎到本站社区发帖提问 参与讨论，获取更多帮助，或者扫码二维码加入 Web 技术交流群。
绑定邮箱获取回复消息
由于您还没有绑定你的真实邮箱，如果其他用户或者作者回复了您的评论，将不能在第一时间通知您！
发布评论
评论（2）
查看器工作站不查询模态，而是查询模态先前推送的图像档案。查看器通过查询/检索 DICOM 服务来查询图像存档。可能有例外，但模式本身不会存档图像。
一种模态通过 SCU/SCP 通信将图像推送到存档，模态是 SCU，对 SCP 进行存档。 Modality 使用存储承诺服务。
还可以给自己买一本像 Oleg S. Pianykh 的这样的书，因为互联网没有真正的友好的 DICOM 学习资源。
A viewer workstation doesn't query a modality, but an image archive previously pushed by the modality. Image archive are queried by a viewer with the query/retrieve DICOM service. There may be exceptions, but modalities don't archive the images themselves.
A modality pushes images to an archive with SCU/SCP communication, modality being the SCU, archive the SCP. Modality uses the Storage commitment service.
Also get yourself a book like Oleg S. Pianykh's, since Internet has no really friendly DICOM learning resources.
每个 PACS 必须支持一些查询键（即搜索键）的规则；当然可能会支持更多。
PACS 中的每个图像（更一般地说，几乎每个对象）都按层次结构排序，因此每个图像都属于一个系列，该系列又属于一个研究，又属于一个患者。在每个级别上，图像都携带描述该级别的信息。例如，在患者级别，图像包含有关患者姓名、出生日期和性别的信息。
查询 PACS 时，客户端必须遵循与该层次结构完全对应的严格查询模型。实际上，至少有 3 个您感兴趣的模型，并且客户端在与 PACS（研究根/患者根/仅患者研究模型）协商连接（关联）时必须选择这些模型。
根据该模型，您必须从最高层开始（通常是上述模型中的患者或研究），依次单独查询不同的层次结构级别。
DICOM 标准的第 4 部分说明了每个模型，您可以查询哪些搜索键（即服务器至少支持什么）以及您必须通过从多个搜索键组成搜索来遵循哪些规则。您必须阅读有关查询/检索的第 4 部分，否则您将无法理解它到底是如何工作的。
匹配键和返回键的术语主要不用于 PACS 查询/检索，而是用于模态工作列表。一般来说，匹配密钥是您发送的密钥以及用于在 PACS 中搜索的值。返回键是发送空的键，即。您要求 PACS 填写您使用匹配键过滤的所有响应，并填写请求的返回键的值。该机制可用于与工作列表中相同的PACS查询——匹配/返回术语仅用于工作列表以更好地描述服务器要求。
哈特哈，
迈克尔
there are some rules which query keys (i.e. search keys) must be supported by every PACS; of course it may support more.
Every image (more general, nearly every object) in the PACS is sorted into a hierarchy, so every image belongs to a Series, which again belongs to a Study which again belongs to a patient. On every of those levels, an image carries information to describe that level. For example on Patient level the image has information about the patient name, birth date and sex.
When querying the PACS, the client has to follow a strict query model that corresponds to exactly that hierarchy. There are actually at least 3 models that are interesting for you and that a client has to select when negotiating a connection (association) with the PACS (Study Root/Patient Root/Patient-study Only Models).
According to the very model, you must query the different hierarchy levels seperately one after another, starting with the highest one (usually patient or study in the above models).
Part 4 of the DICOM standard tells for each of the models, which search keys you might query for (i.e. what the server at LEAST supports) and which rules you have to follow by composing your search out of several search keys. You must read part 4 about query/retrieve, otherwise you won't manage to understand how it exactly works.
The terminology of matching keys and return keys is mostly used with regard not to PACS query/retrieve but to Modality Worklist. In general a matching key is a key that you send with a value to search the PACS for. A return key is a key that send empty, ie. you ask the PACS to fill up all responses that you filtered with the matching keys, to also fill in the value for the requested return keys. The mechanism can be used for PACS query the same as in worklist -- the terminology of matching/return is only used for worklist to better describe the server requirements.
HTH,
Michael