To install build the modnet docker image run the following command inside the modnet folder (the one containing the Dockerfile file).
docker build -t modnet .The above command uses the following switches:
-t container_name: specifies a container name;
To run the Docker image as a daemon, run the following command.
docker run -d -p 127.0.0.1:5000:5000/tcp modnetThe above command uses the following switches:
-d: run the Docker container in daemon mode;-p host:host_port:container_port/proto: bindscontainer_portto thehost_porton the host address using the protocolproto;
To check if the Docker image is running, run the following command.
docker psAnd the output should look like this:
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
132763ced3ab modnet "/opt/app/app.py" 2 seconds ago Up 2 seconds 0.0.0.0:5000->5000/tcp, :::5000->5000/tcp musing_proskuriakova
To stop a container, run the following command:
docker stop {container_id}Where {container_id} is the Container ID that appears in docker ps output. To stop the example container, run docker stop 132763ced3ab.
To extract the background from an image, do the following:
curl -F 'image=@/path/to/image.jpg' http://localhost:5000/put/imageThis will place the image.jpg in the processing queue. The supported image types are image/jpeg and image/png. More can be easily added. The output of the command will produce the following (the token is an example token, expect different tokens for different images):
{"token": "1a4d524ad2c21a7f50dc64ce4ee3a345e28972961c16513465d5161a8c0a3d1b"}Here, the token can be used to retrieve the result. To retrieve the result, use the output of the first command and run the following:
curl -X POST http://localhost:5000/get/json -H 'Content-Type: application/json' -d '{"token": "1a4d524ad2c21a7f50dc64ce4ee3a345e28972961c16513465d5161a8c0a3d1b"}'No need to pass the mimetype. This is stored internally by the container and the output will always follow the same mimetype as the input image. This will produce a response depending on the container. For the segmentation containers, the output json will look like this:
{"url": "/get/image/1a4d524ad2c21a7f50dc64ce4ee3a345e28972961c16513465d5161a8c0a3d1b.jpg", "status": "success"}If the container is busy processing other files, the output.png file will contain the following string {"wait":"true"}. Otherwise, it will be a normal PNG image in which the background was removed.
In order to get the image, just do the curl command on the output URL (don't forget to append the absolute path):
curl http://localhost:5000/get/image/1a4d524ad2c21a7f50dc64ce4ee3a345e28972961c16513465d5161a8c0a3d1b.jpg --output out.jpgRemember that the localhost and the 5000 port are the ones outside from the container. This is because the container is not aware of what runs outside.
You can customize the background via the background parameter:
curl -F 'background=#ffffff' -F 'image=@/path/to/image.jpg' http://localhost:5000/put/imageWill put a white background image behind the detected object. Any 6-hex digit color code is supported. Default is black.
curl -F 'background=https://picsum.photos/200' -F 'image=@/path/to/image.jpg' http://localhost:5000/put/imageWill put a lorem ipsum image behind the detected object.
To detect the objects inside an image, do the following:
curl -F 'image=@/path/to/image.jpg' http://localhost:5000/put/imageThis will place the image.jpg in the processing queue. The supported image types are image/jpeg and image/png. More can be easily added. The output of the command will produce the following:
{"token": "1a4d524ad2c21a7f50dc64ce4ee3a345e28972961c16513465d5161a8c0a3d1b", "mime": "image/jpeg"}Here, the token can be used to retrieve the result by calling the get/json endpoint. To retrieve the result, use the output of the first command and run the following:
curl -X POST http://localhost:5000/get/json -H 'Content-Type: application/json' -d '{"token": "1a4d524ad2c21a7f50dc64ce4ee3a345e28972961c16513465d5161a8c0a3d1b", "mime": "image/jpeg"}' --output /path/to/output.jsonThe result is a JSON response that contains the following keys:
- results: this is an array of detected objects, sorted by the percentage of the image that the image covers; an object has the following keys:
- class: this is the name of the detected object; can be: person bicycle car motorcycle airplane bus train truck boat trafficlight firehydrant stopsign parkingmeter bench bird cat dog horse sheep cow elephant bear zebra giraffe backpack umbrella handbag tie suitcase frisbee skis snowboard sportsball kite baseballbat baseballglove skateboard surfboard tennisracket bottle wineglass cup fork knife spoon bowl banana apple sandwich orange broccoli carrot hotdog pizza donut cake chair couch pottedplant bed diningtable toilet tv laptop mouse remote keyboard cellphone microwave oven toaster sink refrigerator book clock vase scissors teddybear hairdrier toothbrush; the list is available in yolov4/coco.names file;
- conf: this is the confidence level of the object's detection - a number between 0 and 1, 0 = no confidence (actually the threshold is set to 0.5) and 1.0 = 100% sure;
- x: the center of the object's bounding box on the horizontal, starting from the left, in pixels;
- y: the center of the object's bounding box on the vertical, starting from the top, in pixels;
- w: the width of the object's bounding box in pixels;
- h: the height of the object's bounding box in pixels;
- area: the percentage of the image covered by the object - 0 = 0%, 1.0 = 100%
- status: can be one of the following:
- "not queued": the image is not yet placed in the processing queue; in this case, in the JSON output you'll find "wait": "true"
- "processing": the image is in the processing queue, but still being processed; in this case, in the JSON output you'll find "wait": "true"
- "success": the image was successfully processed and the result key is present in the JSON output;
- error: present if an error has happened and it will contain an error message
- "unknown image format": if the image is not PNG or JPG
- "invalid model output": the model has not produced correctly formated JSON data
- "unknown token": the token used to get the output most likely expired (or didn't existed in the first place) and the image requires resubmission;
For ease of testing, you can use the old get/image endpoint to get an image that has the object bounding boxes drawn on top of the original image, to check that the detection works.
- if the token has expired, you will receive a 400 HTTP status code and the response body will be '{"error":"unknown token"}';
- if the image type is not supported (png or jpg) a 400 HTTP status code is returned and the response body will be '{"error":"unknown image format"}';
To start the container in the debug mode, run the following command:
docker run -p 127.0.0.1:5000:5000/tcp -it --entrypoint=/bin/bash modnetThis will get you a bash to the inside of the container, but will not run the Python app. The switches used are:
-it: this is actually-i -t,-istarting the container in interactive mode and-tgranting it a tty for starting a console;--entrypoint=/bin/bash: when a container starts, it executes by default a command (in this case/opt/app/app.py); this will override the command and will actually run the/bin/bash;
To run the Python app inside the container, run:
/opt/app/app.pyAiDocker copyright belongs to Asociatia Pro-Youth, a Romanian not-for-profit organization established under Romanian's Government Emergency Ordinance 26/2000 amended. Having full ownership over this repository, Pro-Youth decided to release AiDocker as a free software under the GNU General Public License, version 3 as published by the Free Software Foundation: you can redistribute it and/or modify it under the terms of said license.
AiDocker is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.