T IModelPrediction.FetchOutput <T, U>(U inputTex)
{
//////// TODO : shape should be passed as argument to Transform Input
var shape = new TFShape(1, inputWidth, inputHeight, 3);
////////
var scaled = TextureTools.scaled(inputTex as Texture2D, 224, 224, FilterMode.Trilinear);
var rotated = TextureTools.RotateImageMatrix(scaled.GetPixels32(), scaled.width, scaled.height, 180);
var input = graph[inputName][0];
var inputTensor = TFSharpUtils.TransformInput(rotated, 224, 224, inputMean, inputStd);
var runner = session.GetRunner();
runner.AddInput(input, inputTensor).Fetch(graph[outputName][0]);
var output = runner.Run()[0];
var feautures = output.GetValue() as float[, , , ];
var jagged_features = ((float[][][][])output.GetValue(jagged: true));
var flat3D = jagged_features.SelectMany(a => a).ToArray();
var flat2D = flat3D.SelectMany(a => a).ToArray();
var flat1D = flat2D.SelectMany(a => a).ToArray();
List <float> featureVector = new List <float>(flat1D);
inputTensor.Dispose();
output.Dispose();
return(featureVector as T);
}
/// <summary>
/// This is the model inference method. You specify an input and the relevant output
/// of the model is extracted.
/// </summary>
/// <typeparam name="T"></typeparam>
/// <typeparam name="U"></typeparam>
/// <param name="inputTex"></param>
/// <returns></returns>
public T FetchOutput <T, U>(U inputTex) where T : class, IList where U : class
{
//////// TODO : shape should be passed as argument to Transform Input
//var shape = new TFShape(1, inputWidth, inputHeight, 3);
////////
//var scaled = TextureTools.scaled(inputTex as Texture2D, 224, 224, FilterMode.Trilinear);
var scaled = GenericUtils.Resize(inputTex as Texture2D, 224, 224);
//var rotated = TextureTools.RotateImageMatrix(scaled.GetPixels32(), scaled.width, scaled.height, 180);
var input = graph[inputName][0];
//var inputTensor = TFSharpUtils.TransformInput(rotated, 224, 224, inputMean, inputStd);
var inputTensor = TFSharpUtils.TransformInput(scaled.GetPixels32(), 224, 224, inputMean, inputStd);
var runner = session.GetRunner();
runner.AddInput(input, inputTensor).Fetch(graph[outputName][0]);
var output = runner.Run()[0];
var outputs = output.GetValue() as float[, ];
inputTensor.Dispose();
output.Dispose();
UnityEngine.Object.DestroyImmediate(scaled);
var list = new List <KeyValuePair <string, float> >();
for (int i = 0; i < labels.Length; i++)
{
var confidence = outputs[0, i];
if (confidence < 0.01f)
{
continue;
}
list.Add(new KeyValuePair <string, float>(labels[i], confidence));
}
var results = list.OrderByDescending(i => i.Value).Take(numOfResults).ToList();
return(results as T);
}
