Spaces:
Runtime error
Runtime error
File size: 1,925 Bytes
86756d8 |
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 |
"""
This module allows to extract texts from videos using OCR
"""
import easyocr
import os
import cv2
import shutil
import difflib
import re
from tools.video_tools import generate_frames
CONF_THRESH = 0.9
SIMILARITY_THRESH = 0.8
def process_text(text):
result = re.sub(r"[\n\"\[\]~;]", "", text)
lst = result.split()
s = ""
for item in lst:
item = item.strip()
if len(item)!=1 or item == "a" or item == "I" or item == "i" or item == "A":
s += " "+item
if len(s)<6:
s = ""
return s
def get_formated_text(texts_arr):
res = ""
for row in texts_arr:
k = process_text(row.lower())
if len(k) > 0:
res += process_text(row.lower()) + ", "
return res[:-2]
def add_text(text_lst, text):
for t in text_lst:
similarity = difflib.SequenceMatcher(None, t, text).ratio()
if similarity > SIMILARITY_THRESH:
return
text_lst.append(text)
def retrieve_text(video_path, rate = 5, frames_path = "tmp_frames", show_print = True):
texts_lst = []
generate_frames(video_path, frames_path, rate = rate, show_print = show_print)
ocr = easyocr.Reader(['en'])
for i in os.listdir(frames_path):
text = ocr.readtext(frames_path + "/" + i)
for txt in text:
# Threshold for confidence
if txt[2] > CONF_THRESH:
# Filter similar texts
add_text(texts_lst, txt[1])
# Delete temporary directory
shutil.rmtree(frames_path)
return texts_lst
def retrieve_to_file(dest, video_path):
text_lst = retrieve_text(video_path, rate = 2, show_print = False)
file = open(dest, "w")
file.writelines([line + "\n" for line in text_lst])
file.close()
def retrieve_to_files(dest, video_path):
for file_name in os.listdir(video_path):
retrieve_to_file(dest + "/" + os.path.splitext(file_name)[0] + "_text.txt", video_path + "/" + file_name)
|