Object Detection

The object detection API locates and classifies 80 different kinds of objects in a single image.

To use this API, you need to set VISION-DETECTION=True when starting DeepStack

sudo docker run -e VISION-DETECTION=True -v localstorage:/datastore \
-p 80:5000 deepquestai/deepstack

If using the GPU Version, run

sudo docker run --rm --runtime=nvidia -e VISION-DETECTION=True -v localstorage:/datastore \
-p 80:5000 deepquestai/deepstack:gpu

Note also that you can have multiple endpoints activated, for example, both face and object detection are activated below

sudo docker run -e VISION-DETECTION=True  -e VISION-FACE=True -v localstorage:/datastore \
-p 80:5000 deepquestai/deepstack

Example

import requests

image_data = open("test-image3.jpg","rb").read()

response = requests.post("http://localhost:80/v1/vision/detection",files={"image":image_data}).json()

for object in response["predictions"]:
    print(object["label"])

print(response)

Result

dog
person
person
{'predictions': [{'x_max': 819, 'x_min': 633, 'y_min': 354, 'confidence': 99, 'label': 'dog', 'y_max': 546}, {'x_max': 601, 'x_min': 440, 'y_min': 116, 'confidence': 99, 'label': 'person', 'y_max': 516}, {'x_max': 445, 'x_min': 295, 'y_min': 84, 'confidence': 99, 'label': 'person', 'y_max': 514}], 'success': True}

We can use the coordinates returned to extract the objects

import requests
from PIL import Image

image_data = open("test-image3.jpg","rb").read()
image = Image.open("test-image3.jpg").convert("RGB")

response = requests.post("http://localhost:80/v1/vision/detection",files={"image":image_data}).json()
i = 0
for object in response["predictions"]:

    label = object["label"]
    y_max = int(object["y_max"])
    y_min = int(object["y_min"])
    x_max = int(object["x_max"])
    x_min = int(object["x_min"])
    cropped = image.crop((x_min,y_min,x_max,y_max))
    cropped.save("image{}_{}.jpg".format(i,label))

    i += 1

Result

Performance

DeepStack offers three modes allowing you to tradeoff speed for peformance. During startup, you can specify performance mode to be , “High” , “Medium” and “Low”

The default mode is “Medium”

You can speciy a different mode as seen below

sudo docker run -e MODE=High -e VISION-DETECTION=True -v localstorage:/datastore \
-p 80:5000 deepquestai/deepstack

Note the -e MODE=High above

Setting Minimum Confidence

By default, the minimum confidence for detecting objects is 0.45. The confidence ranges between 0 and 1. If the confidence level for an object falls below the min_confidence, no object is detected.

The min_confidence parameter allows you to increase or reduce the minimum confidence.

We lower the confidence allowed below.

Example

import requests

image_data = open("test-image3.jpg","rb").read()

response = requests.post("http://localhost:80/v1/vision/detection",
files={"image":image_data},data={"min_confidence":0.30}).json()

CLASSES

The following are the classes of objects DeepStack can detect in images

person,   bicycle,   car,   motorcycle,   airplane,
bus,   train,   truck,   boat,   traffic light,   fire hydrant,   stop_sign,
parking meter,   bench,   bird,   cat,   dog,   horse,   sheep,   cow,   elephant,
bear,   zebra, giraffe,   backpack,   umbrella,   handbag,   tie,   suitcase,
frisbee,   skis,   snowboard, sports ball,   kite,   baseball bat,   baseball glove,
skateboard,   surfboard,   tennis racket, bottle,   wine glass,   cup,   fork,
knife,   spoon,   bowl,   banana,   apple,   sandwich,   orange, broccoli,   carrot,
hot dog,   pizza,   donot,   cake,   chair,   couch,   potted plant,   bed, dining table,
toilet,   tv,   laptop,   mouse,   remote,   keyboard,   cell phone,   microwave,
oven,   toaster,   sink,   refrigerator,   book,   clock,   vase,   scissors,   teddy bear,
hair dryer, toothbrush.