VIAME

Video and Image Analytics for Multiple Environments.

Originally made for marine life detection, VIAME has evolved into a general open-source computer vision toolkit for analysing imagery and video: detection, tracking, classification, size measurement, stereo/multi-camera processing, search and training, assembled into configurable pipelines.

This wheel carries the VIAME runtime and the viame command line tool.

Install

pip install viame

Bring your own GPU stack: torch comes with the CUDA runtime this build links against, and nothing CUDA is bundled here.

Use

Every tool is a subcommand of viame. viame help lists them with a one line description each, and viame help <tool> prints that tool's options.

viame help                                       # list every tool
viame run my_pipeline.pipe                       # run a single pipeline file as-is
viame run my_pipeline.pipe video.mp4             # run a pipeline on a video
viame run detector.zip video.mp4                 # run a detector on a video
viame run detector.zip image_list.txt            # ... or a list of images
viame run detector.pipe videos/                  # run a pipeline over a folder
viame train data/ train_detector.conf            # train a model on some data
viame score detections.csv groundtruth.csv       # score file against groundtruth
viame score computed/ groundtruth/               # score folder against groundtruth
viame csv detections.csv --print-types           # inspect a VIAME csv
viame json tracks.json --print-types             # inspect a DIVE or COCO json
viame add-ons                                    # list and install model add-ons

import viame

Models

The pipelines that ship are those whose models are small enough to travel with the code. Larger models are add-on packs, fetched at runtime into the environment alongside the shipped pipelines:

viame add-ons

Links

• viametoolkit.org
• Source
• Documentation

BSD 3-Clause licensed.