Package biopb.image

Interface ObjectDetectionGrpc.AsyncService

All Known Implementing Classes:
ObjectDetectionGrpc.ObjectDetectionImplBase
Enclosing class:
ObjectDetectionGrpc
public static interface ObjectDetectionGrpc.AsyncService
 Implementing the `RunDetection` call is mandatory. The rest are optional.

  • Method Summary

    Modifier and Type
    Method
    Description
    default void
    getOpNames(com.google.protobuf.Empty request, io.grpc.stub.StreamObserver<OpNames> responseObserver)
    Query supported model names and their kwargs schemas.
    default void
    runDetection(DetectionRequest request, io.grpc.stub.StreamObserver<DetectionResponse> responseObserver)
    Unitary call for computing cell detection / segmentation.
    default io.grpc.stub.StreamObserver<DetectionRequest>
    runDetectionOnGrid(io.grpc.stub.StreamObserver<DetectionResponse> responseObserver)
    Send the image dataset as a stream of (overlapping or non-overlapping) image patches, instead of one large chunk of binary data.
    default io.grpc.stub.StreamObserver<DetectionRequest>
    runDetectionStream(io.grpc.stub.StreamObserver<DetectionResponse> responseObserver)
    The streaming version allows dynamic updating of the DetectionRequest and getting updated results.
    default io.grpc.stub.StreamObserver<DetectionRequest>
    runModelAdaptation(io.grpc.stub.StreamObserver<com.google.protobuf.Any> responseObserver)
    Run model adaptation based on the submitted stream of DetectionRequest, which may or may not contain ROI annotations.

  • Method Details

    • runDetection

      default void runDetection(DetectionRequest request, io.grpc.stub.StreamObserver<DetectionResponse> responseObserver)
       Unitary call for computing cell detection / segmentation.

    • runDetectionStream

      default io.grpc.stub.StreamObserver<DetectionRequest> runDetectionStream(io.grpc.stub.StreamObserver<DetectionResponse> responseObserver)
       The streaming version allows dynamic updating of the DetectionRequest and
 getting updated results. For example, a client may request analyses at
 several different settings on the same image. After initial request, the
 following streaming requests no longer need to transmit the image pixel
 data anymore, but only the new parameter settings, which saves network
 bandwidth.

    • runDetectionOnGrid

      default io.grpc.stub.StreamObserver<DetectionRequest> runDetectionOnGrid(io.grpc.stub.StreamObserver<DetectionResponse> responseObserver)
       Send the image dataset as a stream of (overlapping or non-overlapping) image
 patches, instead of one large chunk of binary data. There are two use
 cases for this call: (1) The image dataset is too large therefore sending
 all data in one message would exceed GRPC message size limit; (2) The
 original data were taken as a grid scan so it is convenient to read/write
 as patches.

    • runModelAdaptation

      default io.grpc.stub.StreamObserver<DetectionRequest> runModelAdaptation(io.grpc.stub.StreamObserver<com.google.protobuf.Any> responseObserver)
       Run model adaptation based on the submitted stream of DetectionRequest,
 which may or may not contain ROI annotations. This call returns an opaque
 data structure representing the model configuration after model
 adaptation, which in turn should be included in future inference request
 as a DetectionSettings field.

    • getOpNames

      default void getOpNames(com.google.protobuf.Empty request, io.grpc.stub.StreamObserver<OpNames> responseObserver)
       Query supported model names and their kwargs schemas.