thot-detector/not_thot_downloader.py
2022-08-01 19:39:28 -05:00

39 lines
1.3 KiB
Python

import urllib.request
import cv2
import numpy as np
import os
def store_raw_images():
blank_img = cv2.imread("thots_false/blank.jpg", cv2.IMREAD_GRAYSCALE)
#thots
#lnk = "http://www.image-net.org/api/text/imagenet.synset.geturls?wnid=n09972458"
#not thots
#lnk = "http://www.image-net.org/api/text/imagenet.synset.geturls?wnid=n00017222"
lnk = "http://www.image-net.org/api/text/imagenet.synset.geturls?wnid=n04081281"
neg_image_urls = urllib.request.urlopen(lnk).read().decode()
n = 1273
for image_url in neg_image_urls.split("\n"):
try:
n += 1
if n >= 2000:
break
path = "thots_false/" + str(n) + ".jpg"
print("\n")
print(str(n) + "/" + str(len(neg_image_urls.split("\n"))))
print(image_url)
urllib.request.urlretrieve(image_url, path)
img = cv2.imread(path, cv2.IMREAD_GRAYSCALE)
if (img.shape == blank_img.shape):
print("Blank Image Detected, deleting...")
os.remove(path)
continue
rs_img = cv2.resize(img, (100, 100))
cv2.imwrite("thots_false/" + str(n) + ".jpg", rs_img)
print("Success")
except Exception as e:
print(str(e))
store_raw_images()