Django REST framework API for object detection. Model trained on COCO dataset.
- Clone the repository:
git clone https://github.com/robot-le/object_detection.git
- Build a docker image:
docker build . --tag object_detection
- Run the container:
docker run -p 8084:8000 object_detection
- Send a request with such tools as Postman or Insomnia, or send it from a terminal with curl:
curl --location 'http://localhost:8084/api/detect-objects/' --form 'image=@"<path_to_your_image>"'
In Postman send POST request with this photo in body as form-data:
In response you'll get an object with a list of detected objects with label, confidence, and bounding box coordinates:
{
"detected_objects": [
{
"label": "car",
"confidence": 0.9929202198982239,
"box": {
"a": 70,
"b": 344,
"c": 266,
"d": 394
}
},
{
"label": "person",
"confidence": 0.9896812438964844,
"box": {
"a": 266,
"b": 330,
"c": 288,
"d": 376
}
},
{
"label": "person",
"confidence": 0.9634418487548828,
"box": {
"a": 295,
"b": 332,
"c": 311,
"d": 376
}
},
{
"label": "person",
"confidence": 0.9565578699111938,
"box": {
"a": 416,
"b": 328,
"c": 440,
"d": 378
}
},
{
"label": "person",
"confidence": 0.9295955300331116,
"box": {
"a": 588,
"b": 326,
"c": 620,
"d": 380
}
},
{
"label": "person",
"confidence": 0.9230886101722717,
"box": {
"a": 306,
"b": 331,
"c": 320,
"d": 375
}
},
{
"label": "person",
"confidence": 0.9230819344520569,
"box": {
"a": 625,
"b": 326,
"c": 639,
"d": 368
}
},
{
"label": "traffic light",
"confidence": 0.9148918390274048,
"box": {
"a": 20,
"b": 163,
"c": 44,
"d": 201
}
},
{
"label": "person",
"confidence": 0.9024598598480225,
"box": {
"a": 356,
"b": 333,
"c": 372,
"d": 375
}
},
{
"label": "car",
"confidence": 0.8786023855209351,
"box": {
"a": 371,
"b": 334,
"c": 401,
"d": 354
}
},
{
"label": "person",
"confidence": 0.8716214299201965,
"box": {
"a": 402,
"b": 327,
"c": 416,
"d": 367
}
},
{
"label": "traffic light",
"confidence": 0.8687766194343567,
"box": {
"a": 440,
"b": 230,
"c": 462,
"d": 252
}
},
{
"label": "person",
"confidence": 0.8665909171104431,
"box": {
"a": 473,
"b": 330,
"c": 495,
"d": 374
}
},
{
"label": "traffic light",
"confidence": 0.7571566700935364,
"box": {
"a": 454,
"b": 232,
"c": 466,
"d": 250
}
},
{
"label": "traffic light",
"confidence": 0.700567901134491,
"box": {
"a": 0,
"b": 164,
"c": 14,
"d": 202
}
},
{
"label": "traffic light",
"confidence": 0.558958888053894,
"box": {
"a": 248,
"b": 300,
"c": 254,
"d": 308
}
},
{
"label": "person",
"confidence": 0.5131950378417969,
"box": {
"a": 120,
"b": 333,
"c": 130,
"d": 357
}
}
]
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent