AWS Documentation » AWS DeepLens » Developer Guide » AWS DeepLens Device Library » awscam Module for Inference » Model Object » model.parseResult Method

model.parseResult Method

Parses the results of some commonly used models, such as classification, SSD, and segmentation models. For customized models, you need to write your own parse functions.

Request Syntax

import awscam
model = awscam.Model(model_topology_file, loading_config)
raw_infer_result = model.doInference(video_frame)
result = model.parseResult(model_type, raw_infer_result)

Parameters

• model_type—String that identifies the model type to use to generate the inference. Required.

  Valid values: classification, ssd, and segmentation

  Note

  When deploying an Amazon SageMaker-trained SSD model, you must first run deploy.py (available from https://github.com/apache/incubator-mxnet/tree/master/example/ssd/) to convert the model artifact into a deployable mode.

• raw_infer_result—The output of the function model.doInference(video_frame). Required.

Return Type

• dict

Returns

Returns a dict with a single key-value pair. The key is the model type. The value is a list of dicts, in which each element is an object label and probability calculated by the model.

The output of a classification model might look like the following:

{
    "output":[
        {"label":"318", "prob":0.5},
        {"label":"277", "prob":0.3},
        ...,
        {"label":"433", "prob":0.001}
    ]
}

The output of an SDD model contains bounding box information, similar to the following:

{
    "output": [
        {"label": "318", "xmin": 124, "xmax": 245, "ymin": 10,  "ymax": 142, "prob": 0.5},
        {"label": "277", "xmin": 89,  "xmax": 166, "ymin": 233, "ymax": 376, "prob": 0.3},
        ...,
        {"label": "433", "xmin": 355, "xmax": 468, "ymin": 210, "ymax": 266, "prob": 0.001}
    ]
}