How to convert MultiArray to image as input type

[MarcSteven]

❓Question

How to convert MultiArray to Image as input type

System Information

• If applicable

[AnasArram]

❓Question

How to convert MultiArray to Image as input type

System Information

• If applicable

you need to pass the input name (image_input_names) as an argument as follow:

coreml_model = convert(model, target_ios='13', custom_conversion_functions={'Upsample': _convert_upsample}, image_input_names = ['input']

replace 'input' with your model input name

[MarcSteven]

@AnasArram Thank you so much, can you show a whole example about it?

[MarcSteven]

`#//
#// convert_input_to_image.py
#// AntsFaceDemo
#//
#// Created by Marc Zhao on 2019/12/12.
#// Copyright © 2019 AntsNetwork. All rights reserved.
#//

import coremltools
import coremltools.proto.FeatureTypes_pb2 as ft

spec = coremltools.utils.load_spec("FacialExpressionAnalysis.mlmodel") #please change the file name you are using

input = spec.description.input[0]
input.type.imageType.colorSpace = ft.ImageFeatureType.RGB
input.type.imageType.height = 224
input.type.imageType.width = 224

coremltools.utils.save_spec(spec,"FacialExpressionAnalysis.mlmodel")#Please change the file name using your own
`
This is all the solutions, but you must ensure the channel is 3, you can convert multiArray to image if you want.

[yonggunlee]

Hello, I don't know if this is still on going or not, but I have a question.
I see how we can change input&output type in coremltools.
I wonder if it can be happen in onnxmltools as well.
I have onnx model that takes [batchsize, img channel, img H, img W] which is multi array,
I also wish to change the input as image input as above question.
Can someone help me on it? I was trying to check inside of onnxmltools, but haven't found how.