Visual Intelligence Platform

Relational data : Postgres, MYSQL, SQLite


Text, HTML : Lucene/Solr, Elasticsearch


Videos & Images : Deep Video Analytics


Upload videos or set of images. Provide Youtube urls for automatic downloading. Browse & annotate uploaded videos. Ability to import pre-indexed datasets coming soon.


Perform scene detection, frame extraction on videos. Detect objects in frames/images, index entire frames/images, detected objects using deep learning algorithms.


Extracted objects, along with entire frames and crops, are indexed using deep features. The feature vectors are used for retrieval by query image.


Using Docker deploy on variety of machines with/without GPUs, local and cloud. Docker compose enables automated setup including database & RabbitMQ.

Extensible platform

Metadata & files

Metadata is stored in Postgres for easy extensibility. Files and vectors are stored in a consistent schema within a single folder.


Async queues

Asynchronous processing via celery allows extraction, detection, indexing & query flows to be easily modified.

Features & Models

We take significant efforts to ensure that following models (code+weights included) work without having to write any code.


  • Visual Search using Nearest Neighbors algorithm as a primary interface

  • Upload videos, multiple images (zip file with folder names as labels)

  • Provide Youtube url to be automatically processed/downloaded by youtube-dl

  • Leverage pre-trained object recognition/detection, face recognition models for analysis and visual search.

  • Query against pre-indexed external datasets containing millions of images.

  • Metadata stored in Postgres, Operations performed asynchronously using celery tasks.

  • Separate queues and workers for selection of machines with different specifications (GPU vs RAM).

  • Videos, frames, indexes, numpy vectors stored in media directory, served through nginx

  • Explore data, manually run code & tasks without UI via a jupyter notebook explore.ipynb


Planned external datasets & Models (coming soon!)


Pre-built docker images for both CPU & GPU versions are available on Docker Hub.

Machines without an Nvidia GPU

Deep Video analytics is implemented using Docker and works on Mac, Windows and Linux. Make sure you have latest version of Docker installed.

git clone
cd DeepVideoAnalytics/docker && docker-compose up

Machines with Nvidia GPU

You need to have latest version of Docker and nvidia-docker installed. The GPU Dockerfile is slightly different from the CPU version dockerfile.

pip install --upgrade nvidia-docker-compose
git clone
cd DeepVideoAnalytics/docker_GPU && ./
nvidia-docker-compose up

On AWS using a P2.xlarge instance

We provide an AMI with all dependencies such as docker & nvidia drivers pre-installed. To use it start a P2.xlarge instance with AMI ID: ami-848f3d92 (N. Virginia) and ports 8000, 6006, 8888 open (preferably to only your IP). Run following commands after logging into the machine via SSH. After approximately 5 ~ 1 minutes the user interface will appear on port 8000 of the instance ip. AMI creation is documented here.

cd deepvideoanalytics && git pull
cd docker_GPU && rm nvidia-docker-compose.yml && ./ && nvidia-docker-compose up

Security warning: The current GPU container uses nginx <-> uwsgi <-> django setup to ensure smooth playback of videos. However it runs nginix as root (within the container). Since you can modify AWS Security rules on-the-fly, allow inbound traffic only from your own IP address.


Following options can be specified in docker-compose.yml, or your environment to selectively enable/disable algorithms.

Specify following options via environment variables, to enable additional algorithms.

  • ALEX_ENABLE=1 (to use Alexnet with PyTorch. disabled by default)

  • YOLO_ENABLE=1 (to use YOLO 9000. disabled by default)

Specify following options via environment variables, to modify video processing.

  • SCENEDETECT_DISABLE=1 (to disable scene detection. enabled by default)

  • RESCALE_DISABLE=1 (to disable rescaling of frame extracted from videos. enabled by default)


Docker containers, networking and volumes

Video & Query processing

Documentation & Presentation

Some documentation is available here along with a board for planned future tasks.

For a quick overview of design choices and vision behind this project we strongly recommend going through following presentation.

Paper & Citation

Coming Soon!

Issues, Questions & Contact

Please submit all software related bugs and questions using Github issues, for other questions you can contact me at

© 2017 Akshay Bhat, Cornell University.
All rights reserved.