|
import streamlit as st |
|
import pandas as pd |
|
import numpy as np |
|
import pickle |
|
import base64 |
|
import seaborn as sns |
|
import matplotlib.pyplot as plt |
|
|
|
st.write(""" |
|
# KneeMRI Detection App |
|
|
|
MRI (Magnetic Resonance Imaging) lutut. tes pencitraan diagnostik non |
|
invasif yang dilakukan untuk menganalisis kondisi berbagai bagian lutut, |
|
seperti jaringan, tendon, sendi, otot, tulang, dan ligamen. |
|
|
|
Aplikasi ini mendeteksi kemungkinan terjadinya cedera Robekan Anterior Cruciate Ligament (ACL) |
|
yang biasa dialami oleh atlet papan atas dalam sepak bola atau bola basket. |
|
Cedera ini juga kerap timbul pada usia 40 tahun keatas. |
|
|
|
""") |
|
|
|
url_dataset = f'<a href="kneeMRI.csv">Download Dataset CSV File</a>' |
|
st.markdown(url_dataset, unsafe_allow_html=True) |
|
|
|
def user_input_features() : |
|
kneeLR = st.sidebar.selectbox('Knee', ('Left', 'Right')) |
|
roiX = st.sidebar.slider('RoiX', 71,146) |
|
roiY = st.sidebar.slider('RoiY', 22,184) |
|
roiZ = st.sidebar.slider('RoiZ', 9,22) |
|
roiHeight = st.sidebar.slider('RoiHeight', 62,124) |
|
roiWidth = st.sidebar.slider('RoiWidth', 61,136) |
|
roiDepth = st.sidebar.slider('RoiDepth ', 2,6) |
|
kneeLR01 = 1 |
|
if(kneeLR == 'Left') : |
|
kneeLR01 = 0 |
|
data = {'kneeLR':[kneeLR01], |
|
'roiX':[roiX], |
|
'roiY':[roiY], |
|
'roiZ':[roiZ], |
|
'roiHeight':[roiHeight], |
|
'roiWidth':[roiWidth], |
|
'roiDepth':[roiDepth],} |
|
|
|
features = pd.DataFrame(data) |
|
return features |
|
|
|
input_df = user_input_features() |
|
|
|
knee_raw = pd.read_csv('kneeMRI.csv') |
|
knee_raw.fillna(0, inplace=True) |
|
knee = knee_raw.drop(columns=['aclDiagnosis']) |
|
df = pd.concat([input_df, knee],axis=0) |
|
|
|
df = df[:1] |
|
df.fillna(0, inplace=True) |
|
|
|
features = ['kneeLR', 'roiX', 'roiY', 'roiZ', 'roiHeight', |
|
'roiWidth', 'roiDepth'] |
|
|
|
df = df[features] |
|
|
|
st.subheader('User Input features') |
|
st.write(df) |
|
load_clf = pickle.load(open('kneeMRI_clf.pkl', 'rb')) |
|
detection = load_clf.predict(df) |
|
detection_proba = load_clf.predict_proba(df) |
|
knee_labels = np.array(['Normal','Robek']) |
|
st.subheader('Detection') |
|
st.write(knee_labels[detection]) |
|
st.subheader('Detection Probability') |
|
df_prob = pd.DataFrame(data=detection_proba, index=['Probability'], columns=knee_labels) |
|
st.write(df_prob) |
