Premier-eye
Program for detecting objects
Premier-eye required to create the data and output folder on the device before starting work, and place the images there for recognition work.
Module for detecting objects
Requirements for local use:
- Python >= 3.6
For run module you need run API & SPA for send data or set variable sendRequestToServer
in Premier-eye.DS/config/config.ini to false
Docker use:
cd Premier-eye.DS
docker-compose up --build
On windows, gpu in docker not implemented now
On linux, use nvidia-docker
API & SPA
Run main docker-compose file:
docker-compose up --build
Nomeroff-net
For separately use:
docker build -t nn . && docker run -p 5051:5051 nn
Algorithms
- Mask R-CNN
- Implementation in neural_network/mask
- Example of work
Information for dev-ops
If the connection cannot be established, check with docker network inspect @ netName @
ip address to which it is connected
server, and if necessary, change the parameter API_URL
.