Consistency Live Demo

Update (29/03/2022): For a live demo of the latest models, please visit here where the models are trained using cross-task consistency and 3D data augmentations on Omnidata.

We provide a demo where you can upload (or use an existing) query and see the results of different consistency-based models vs various baselines. The trained models can be downloaded here.

You can also visit the archive page to examine previously uploaded query images by online users and their corresponding results. The image you upload will be added to this archive too, unless you ask us to remove it.

The networks were trained on 3 million images of general indoor scenes (rather than object close-ups, faces, humans, etc). You can see the statistics and samples of the training data here. If your query image severely deviates from the training data, the performance is expected to degrade.

Running the full demo usually takes around 30 seconds, depending on the traffic..

Upload your own image or click on one of the sample queries below. Click on the cube to use a random query image from previous uploads.

sample 1 sample 4 sample 3 random