Development Workflow

Process Description

Figure 1 API development process

Key Steps

1. Initialize Vision SDK APIs before using them to develop applications.
2. Perform media data processing/model inference.
   • Media data processing includes image encoding and decoding, cropping, resizing, padding, and CSC.
   • Model inference, including object recognition and image classification, consists of the following steps.
     1. Before model inference, prepare a proper OM or MindIR model. You can convert a pre-trained model into an OM model.
     2. Load the prepared model to the system as a file or through memory.
     3. (Optional) Perform media data processing. Before model inference, operations can be performed on the input media data, including image decoding, cropping, resizing, and padding.
     4. Run model inference to implement functions such as image classification and object recognition.
     5. (Optional) Perform postprocessing on the data output by model inference. The inference result is processed based on user requirements. For example, it can be a specific result that can be understood and read. Model postprocessing supports the following ways:
        • (Recommended) Select a model postprocessing mode by yourself.
        • Call the Vision SDK APIs.
        • Perform further development based on the existing model.
3. Perform deinitialization by calling MxDeInit(). Perform deinitialization after the entire application process is complete. Otherwise, errors may occur during the subsequent initialization of internal system resources, causing exceptions in other services.