File size: 3,842 Bytes
5eab8c7 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 |
import streamlit as st
import cv2
import numpy as np
import easyocr
import re
from langdetect import detect_langs
from PIL import Image
import io
def load_easyocr_reader():
return easyocr.Reader(['hi', 'en'], gpu=False)
def preprocess_image(image):
img_array = np.array(image.convert('RGB'))
gray = cv2.cvtColor(img_array, cv2.COLOR_RGB2GRAY)
denoised = cv2.fastNlMeansDenoising(gray, None, 10, 7, 21)
thresh = cv2.adaptiveThreshold(denoised, 255, cv2.ADAPTIVE_THRESH_GAUSSIAN_C, cv2.THRESH_BINARY, 11, 2)
kernel = np.ones((1, 1), np.uint8)
dilated = cv2.dilate(thresh, kernel, iterations=1)
return dilated
def perform_easyocr(image, reader):
preprocessed_image = preprocess_image(image)
results = reader.readtext(preprocessed_image, paragraph=True, detail=0,
contrast_ths=0.2, adjust_contrast=0.5,
add_margin=0.1, width_ths=0.7, height_ths=0.7)
extracted_text = ' '.join(results)
return extracted_text
def detect_languages(text):
cleaned_text = re.sub(r'[^a-zA-Z\u0900-\u097F\s]', '', text)
cleaned_text = re.sub(r'\s+', ' ', cleaned_text).strip()
if not cleaned_text:
return []
try:
langs = detect_langs(cleaned_text)
detected = []
for lang in langs:
if lang.lang == 'hi' and lang.prob > 0.1:
detected.append('Hindi')
elif lang.lang == 'en' and lang.prob > 0.1:
detected.append('English')
return detected
except:
return fallback_language_check(cleaned_text)
def fallback_language_check(text):
hindi_range = range(0x0900, 0x097F)
english_range = range(0x0041, 0x007A)
has_hindi = any(ord(char) in hindi_range for char in text)
has_english = any(ord(char) in english_range for char in text)
detected = []
if has_hindi:
detected.append('Hindi')
if has_english:
detected.append('English')
return detected
def highlight_text(text, keywords):
for keyword in keywords:
# Highlight the keyword by wrapping it with a span tag
text = re.sub(f'({re.escape(keyword)})', r'<span style="background-color: yellow;">\1</span>', text, flags=re.IGNORECASE)
return text
def main():
st.title("OCR for Hindi and English")
easyocr_reader = load_easyocr_reader()
# Initialize session state variables
if 'extracted_text' not in st.session_state:
st.session_state.extracted_text = ""
uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])
if uploaded_file is not None:
image = Image.open(uploaded_file)
st.image(image, caption='Uploaded Image', use_column_width=True)
if st.button('Perform OCR'):
with st.spinner('Processing...'):
st.session_state.extracted_text = perform_easyocr(image, easyocr_reader)
st.subheader("Extracted Text:")
st.write(st.session_state.extracted_text)
languages_detected = detect_languages(st.session_state.extracted_text)
if languages_detected:
st.write("Detected languages:", ', '.join(languages_detected))
else:
st.write("No languages detected.")
if st.session_state.extracted_text: # Check if OCR has been performed
st.subheader("Search in Extracted Text")
search_query = st.text_input("Enter keywords to search:", "")
if search_query:
keywords = search_query.split()
highlighted_text = highlight_text(st.session_state.extracted_text, keywords)
st.markdown(highlighted_text, unsafe_allow_html=True)
if __name__ == "__main__":
main() |