def test_batch_image_recognition(self):
test_data = os.path.normpath(os.path.join(DIR, "data", "fear.png"))
response = image_recognition([test_data, test_data], api_key = self.api_key, top_n=3)
self.assertIsInstance(response, list)
self.assertIsInstance(response[0], dict)
self.assertEqual(len(response[0]), 3)
self.assertIsInstance(list(response[0].values())[0], float)
def analyze_image(self, gif_url):
"""
Checks to see if there are any recognizable objects in the gifs, and if there are performs sentiment analysis on the objects. Then it outputs the average of the sentiment analysis of all the objects
"""
gif_output = []
#print(gif_url)
while True:
try:
gif_output = indicoio.image_recognition(gif_url)
break
except ValueError:
#print('oops')
return None
#print(gif_output)
self.list_objects = [
name for name, prob in gif_output.items() if prob > .1
]
if len(self.list_objects) == 1:
self.sentiment_output = indicoio.sentiment_hq(self.list_objects[0])
elif len(self.list_objects) == 0:
return self.text_sentiment
else:
self.sentiment_output = indicoio.sentiment_hq(self.list_objects)
self.sentiment_output = np.mean(self.sentiment_output)
return self.sentiment_output
def processImage():
url = request.args.get("url")
r = requests.get(url, stream = True)
encoded_string = base64.b64encode(r.content)
indicoResult = indicoio.image_recognition(encoded_string, top_n=2, hq=True)
app.logger.debug(indicoResult)
return json.jsonify(uri=encoded_string, keywords=indicoResult)
def test_batch_image_recognition(self):
test_data = os.path.normpath(os.path.join(DIR, "data", "fear.png"))
response = image_recognition([test_data, test_data],
api_key=self.api_key,
top_n=3)
self.assertIsInstance(response, list)
self.assertIsInstance(response[0], dict)
self.assertEqual(len(response[0]), 3)
self.assertIsInstance(list(response[0].values())[0], float)
dresses = ["dress", "gown"]
tops = ["button", "blouse", "shirt", "top"]
bottoms = ["pants", "skirt", "trousers", "denim", "jeans", "shorts", "slacks"]
clothing = []
i=0
for root, dirs, files in os.walk("imgs"):
for image in files:
if image.endswith("jpg"):
clothing.append(os.path.join(root, image))
i+=1
if i == 500:
break
clothes = indicoio.image_recognition(clothing)
j = open('jackets.txt', 'w')
d = open('dresses.txt', 'w')
t = open('tops.txt', 'w')
b = open('bottoms.txt', 'w')
i=0
for cloth in clothes:
sortC = sorted(cloth.items(), key=operator.itemgetter(1), reverse=True)
print(sortC[0][0])
if any(s in sortC[0][0] for s in jackets):
j.write(clothing[i]+"!"+sortC[0][0]+"?")
elif any(s in sortC[0][0] for s in dresses):
d.write(clothing[i]+"!"+sortC[0][0]+"?")
elif any(s in sortC[0][0] for s in tops):
t.write(clothing[i]+"!"+sortC[0][0]+"?")
def test_expected_response(self):
test_data = os.path.normpath(os.path.join(DIR, "data", "keyboard.jpg"))
response = image_recognition(test_data, api_key=self.api_key, top_n=3)
assert "space bar" in response.keys()
def test_expected_response(self):
test_data = os.path.normpath(os.path.join(DIR, "data", "keyboard.jpg"))
response = image_recognition(test_data, api_key = self.api_key, top_n=3)
assert "space bar" in response.keys()
def findThings(thePath):
#finds the objects in the photo
theObj = indicoio.image_recognition(thePath, top_n=3)
return theObj.keys()
import cv2
import time
import indicoio
indicoio.config.api_key = 'd17e09c08d43f673f05743ec7304c9be'
cap = cv2.VideoCapture(0)
while(True):
# Capture frame-by-frame
time.sleep(.2) #a pause so that this doesn't process EVERY frame (twould be overly expensive/time-consuming)
ret, frame = cap.read()
# Display the resulting frame
cv2.imshow('frame',frame)
if cv2.waitKey(1) & 0xFF == ord('q'):
break
highestProb=[0]
highestItem=['thing']
#print indicoio.image_recognition(frame)
probDict= indicoio.image_recognition(frame) #returns a dictionary of 1000 "likely" items and the probability that it is the object...
for item in probDict:
if(probDict[item]>highestProb[0]): #finding the most likely object on the screen
highestProb[0]=probDict[item] #returning a list so that we know what the top 5 objects are
highestItem[0]=item
print highestProb
print highestItem
# When everything done, release the capture
cap.release()
cv2.destroyAllWindows()
