Spaces:
Running
Running
| import streamlit as st | |
| import easyocr | |
| from transformers import AutoModelForTokenClassification, AutoTokenizer, pipeline | |
| from PIL import Image | |
| import numpy as np | |
| import cv2 | |
| # Initialize EasyOCR for text extraction | |
| reader = easyocr.Reader(['en']) | |
| # Load ClinicalBERT for NER (Named Entity Recognition) | |
| model_name = "emilyalsentzer/Bio_ClinicalBERT" | |
| tokenizer = AutoTokenizer.from_pretrained(model_name) | |
| model = AutoModelForTokenClassification.from_pretrained(model_name) | |
| nlp = pipeline("ner", model=model, tokenizer=tokenizer) | |
| # Mock function for validating test results | |
| def validate_medical_value(test_name, test_value): | |
| reference_ranges = { | |
| "Blood Pressure": (90, 120), # Example range: systolic | |
| "Glucose": (70, 100), # Normal range for glucose in mg/dL | |
| } | |
| if test_name in reference_ranges: | |
| min_range, max_range = reference_ranges[test_name] | |
| if test_value < min_range: | |
| return f"{test_name} is low." | |
| elif test_value > max_range: | |
| return f"{test_name} is high." | |
| else: | |
| return f"{test_name} is within the normal range." | |
| else: | |
| return f"{test_name} range is unknown." | |
| # Function to extract text from image | |
| def extract_text_from_image(image): | |
| result = reader.readtext(image) | |
| text = ' '.join([item[1] for item in result]) | |
| return text | |
| # Function to check image clarity (basic approach using edge detection) | |
| def check_image_clarity(image): | |
| # Convert image to grayscale | |
| gray_image = image.convert("L") | |
| image_np = np.array(gray_image) | |
| # Simple edge detection (Laplacian method) | |
| laplacian_var = cv2.Laplacian(image_np, cv2.CV_64F).var() | |
| # Threshold for determining clarity | |
| return laplacian_var > 100 # This threshold can be adjusted | |
| # Function to extract medical entities (test results) from text | |
| def extract_medical_values(text): | |
| entities = nlp(text) | |
| medical_data = {} | |
| for entity in entities: | |
| if entity['entity_group'] in ['LAB_RESULT', 'DISEASE', 'MEDICATION']: # Modify based on actual labels | |
| medical_data[entity['word']] = entity['score'] | |
| return medical_data | |
| # Function to analyze medical report | |
| def analyze_report(image): | |
| # Step 1: Check image clarity | |
| if not check_image_clarity(image): | |
| return "Image is unclear, please upload a clearer image." | |
| # Step 2: Extract text from the medical report image | |
| text = extract_text_from_image(image) | |
| # Step 3: Extract medical values using ClinicalBERT | |
| medical_values = extract_medical_values(text) | |
| analysis_results = [] | |
| for test_name, test_value in medical_values.items(): | |
| # Step 4: Validate test value against normal range | |
| validation_result = validate_medical_value(test_name, test_value) | |
| analysis_results.append(validation_result) | |
| return analysis_results | |
| # Streamlit UI | |
| st.title("Medical Report Analyzer") | |
| st.write("Upload your medical report image to analyze the test results.") | |
| # Upload the image | |
| uploaded_image = st.file_uploader("Upload Image", type=["jpg", "png", "jpeg"]) | |
| if uploaded_image is not None: | |
| # Open the uploaded image | |
| image = Image.open(uploaded_image) | |
| # Show image for verification | |
| st.image(image, caption="Uploaded Medical Report", use_column_width=True) | |
| # Analyze the report | |
| analysis_results = analyze_report(image) | |
| # Display results | |
| if isinstance(analysis_results, list): | |
| for result in analysis_results: | |
| st.write(result) | |
| else: | |
| st.write(analysis_results) | |