devoworm-group/Lineage_Population

1import streamlit as st2import numpy as np3import pandas as pd4from PIL import Image5from pathlib import Path6import joblib7 8import numpy as np9import cv210import onnxruntime as ort11import imutils12# import matplotlib.pyplot as plt13import pandas as pd14import plotly.express as px15 16 17def scale_model_outputs(scaler_path, data):18    scaler= joblib.load(scaler_path)19    scaled=scaler.inverse_transform(data)20    return(scaled)21 22 23def onnx_predict_lineage_population(input_image):24    ort_session = ort.InferenceSession('onnx_models/lineage_population_model.onnx')25    img = Image.fromarray(np.uint8(input_image))26    resized = img.resize((256, 256), Image.NEAREST)27 28    transposed=np.transpose(resized, (2, 1, 0))  29    img_unsqueeze = expand_dims(transposed)30 31    onnx_outputs = ort_session.run(None, {'input': img_unsqueeze.astype('float32')}) 32    return(onnx_outputs[0])33 34 35 36def expand_dims(arr):37    norm=(arr-np.min(arr))/(np.max(arr)-np.min(arr)) #normalize38    ret = np.expand_dims(norm, axis=0)39    return(ret)40 41 42 43def lineage_population_model():44    selected_box2 = st.sidebar.selectbox(45    'Choose Example Input',46    (['Example_1.png'])47    )48 49    st.title('Predict Cell Lineage Populations')50    instructions = """51        Predict the population of cells in C. elegans embryo using fluorescence microscopy data. \n52        Either upload your own image or select from the sidebar to get a preconfigured image. 53        The image you select or upload will be fed through the Deep Neural Network in real-time 54        and the output will be displayed to the screen.55        """56    st.text(instructions)57    file = st.file_uploader('Upload an image or choose an example')58    example_image = Image.open('./images/lineage_population_examples/'+selected_box2).convert("RGB")59 60    col1, col2= st.columns(2)61 62    if file:63        input = Image.open(file).convert("RGB")64        fig1 = px.imshow(input, binary_string=True, labels=dict(x="Input Image"))65        fig1.update(layout_coloraxis_showscale=False)66        fig1.update_layout(margin=dict(l=0, r=0, b=0, t=0))67        col1.plotly_chart(fig1, use_container_width=True)68    else:69        input = example_image70        fig1 = px.imshow(input, binary_string=True, labels=dict(x="Input Image"))71        fig1.update(layout_coloraxis_showscale=False)72        fig1.update_layout(margin=dict(l=0, r=0, b=0, t=0))73        col1.plotly_chart(fig1, use_container_width=True)74 75    pressed = st.button('Run')76    if pressed:77        st.empty()78        output = onnx_predict_lineage_population(np.array(input))79        scaled_output = scale_model_outputs(scaler_path="./scaler.gz", data=output)80 81        for i in range(len(scaled_output[0])):82            scaled_output[0][i]=int(round(scaled_output[0][i]))83 84        df = pd.DataFrame({"Lineage":["A", "E", "M", "P", "C", "D", "Z"] , "Population": scaled_output[0]})85        col2.table(df)