Public | Automated Build

Last pushed: a year ago
Short Description
Real-time camera image recognition.
Full Description

SCRT project

AI for security.

How to run

First you need to install docker on your machine: https://docs.docker.com/engine/installation/linux/ubuntulinux/

To remove local docker image, e.g. scrtai/cam, type:

$ docker rmi scrtai/cam

TensorFlow Serving server

Installing using docker:

$ docker run -d --name=inception -p 9000:9000 scrtai/inception /bin/bash -c "/serving/bazel-bin/tensorflow_serving/model_servers/tensorflow_model_server --port=9000 --model_name=inception --model_base_path=/serving/inception-export &> inception_log"

HTTP API:

Installing using docker:

$ docker run -d --name=inception_api -p 8080:8080 --link=inception:tf -e DEBUG=scrt* scrtai/api

Post JPEG image to localhost:8080

Camera real-time recognition:

Installing using docker:

$ docker run -d --name=inception_cam --link=inception:tf -e POST_SERVER=http://scrt.credo.ru/notify.php -e DEBUG=scrt* --device=/dev/video0:/dev/video0 scrtai/cam

Check container logs:

$ docker logs -f inception_cam
Docker Pull Command
Owner
scrtai
Source Repository

Comments (0)