N2V-Fiji-Demo.imjoy.html

<docs lang="markdown">
## PyImageJ demo

A PyImageJ demo ported from https://github.com/imagej/pyimagej

</docs>

<config lang="json">
{
    "name": "N2V-Fiji-Demo",
    "type": "native-python",
    "version": "0.1.4",
    "api_version": "0.1.2",
    "description": "This plugin process image with Noise2Void plugin in Fiji",
    "tags": [],
    "ui": "",
    "inputs": null,
    "outputs": null,
    "flags": [],
    "icon": "extension",
    "env": ["conda create -n imagej", {"type": "binder", "spec": "oeway/pyimagej-binder-image/master", "skip_requirements": true}],
    "requirements": ["conda: openjdk=8", "conda: -c conda-forge pyimagej", "pip: scikit-image pillow"],
    "dependencies": []
}
</config>

<script lang="python">
from imjoy import api
import imagej
import numpy as np
import skimage
from skimage import io
import numpy as np
from PIL import Image
import base64
from io import BytesIO
import urllib.request

import os
ij = imagej.init(os.path.expanduser("~") + '/Fiji.app',headless=True)

    
def array2base64(img):
    img = img/(img.max())*255.0
    img = Image.fromarray(img.astype('uint8'))
    byte_io = BytesIO()
    img.save(byte_io, 'PNG')
    result = base64.b64encode(byte_io.getvalue()).decode('ascii')
    imgurl = 'data:image/png;base64,' + result
    return imgurl

class ImJoyPlugin():
    def setup(self):
        api.log(ij.getVersion().toString())
        urllib.request.urlretrieve('https://github.com/juglab/N2V_fiji/blob/master/src/test/resources/format-0.2.0-csbdeep.bioimage.io.zip?raw=true', '/home/jovyan/format-0.2.0-csbdeep.bioimage.io.zip')

    async def run(self, my):
        # Import an image with scikit-image.
        # NB: Blood vessel image from: https://www.fi.edu/heart/blood-vessels
        await api.showStatus('Loading example image...')
        urllib.request.urlretrieve('https://raw.githubusercontent.com/juglab/N2V_fiji/master/src/test/resources/blobs.tif', '/home/jovyan/input.tif')
        
        api.showStatus('Processing image with N2V...')
        
        modelPath = "/home/jovyan/format-0.2.0-csbdeep.bioimage.io.zip"
        inputPath = "/home/jovyan/input.tif"
        outputPath = "/home/jovyan/output.tif"
        if os.path.exists(outputPath):
            os.remove(outputPath)
        args = { 
            'training': '/home/jovyan/input.tif',
            'prediction': '/home/jovyan/input.tif',
            'output': outputPath,
            'axes': 'XY',
            'mode3d': False,
            'numepochs': 2,
            'numstepsperepoch': 4,
            'batchsize': 8,
            'patchshape': 16,
            'neighborhoodradius': 5
        }
        modelzoo_service = ij.context().getService('net.imagej.modelzoo.ModelZooService')
        model = modelzoo_service.io().open(modelPath)
        img = ij.io().open(inputPath)
        output = modelzoo_service.predict(model, img, args['axes'])
        dataset = ij.dataset().create(output.get('output'))
        ij.io().save(dataset, outputPath)
        
        await api.showStatus('Image processed with ImageJ.')

        img = io.imread('/home/jovyan/input.tif')
        output_img = io.imread('/home/jovyan/output.tif')
        # Display image
        imgurl1 = array2base64(img)
        imgurl2 = array2base64(output_img)
        viewer = await api.createWindow(src="https://kaibu.org")
        await viewer.view_image(imgurl1, name="input image")
        await viewer.view_image(imgurl2, name="output image")
        # api.showDialog({'type': 'imjoy/image-compare', 'name': 'ImageJ Demo', 'data': {"first": imgurl1, 'second': imgurl2, 'name': 'Process with frangiVesselness in ImageJ'}})

api.export(ImJoyPlugin())
</script>