SketchQL is a video database management system for retrieving video moments with a sketch-based query interface.This interface allows users to specify object trajectory events with simple mouse drag-and-drop operations. Using a pre-trained model that encodes trajectory similarity, SketchQL achieves zero-shot video moments retrieval by performing similarity searches over the video to identify clips that are the most similar to the visual query.
-
First, set up the backend following the SketchQL installation guide.
-
Start the backend server
# Navigate to SketchQL/
python3 server.py
- Clone the
tldraw
repository. Please clone v1 (original version) instead of the new tldraw.
git clone https://github.com/tldraw/tldraw-v1.git
-
Replace the
tldraw-v1/examples/core-example-advanced
folder with thecore-example-advanced
folder from this repository -
Install dependencies
# Install yarn if not already installed
npm install -g yarn
# Navigate to project and install dependencies
cd tldraw-v1
yarn install
- Start the application
yarn start:core
- Open your browser and navigate to:
http://localhost:5421
. If you are seeing a blank canvas on Chrome, try Safari instead.
A video demonstrating how SketchQL works can be found under the video/
folder
- Backend Repository: Georgia Tech DB SketchQL
- tldraw Framework: tldraw-v1 GitHub