FAQs

Everything you need to know about Art Recognition all in one place.

Browse our frequently asked questions to quickly find out if we have already addressed your query.

If we have not managed to answer your query, then please email us as at office@art-recognition.com so that we can provide the needed information.

What factors influence the price of an AI authenticity evaluation?

The price depends on how difficult it is to get enough images, or whether there is disagreement among experts as to which paintings are to be regarded as authentic. Ultimately, it is also decisive how often an artist has changed his style, whether there have been a lot of restorations or whether an artist had his apprentices finish an artwork.

I have an artwork with an unknown author. Should I contact Art Recognition nevertheless?

Definitely. Please also provide additional information such as the style, period, and a short list containing the most probable candidates.

Is Art Recognition specialized on a particular artist or group of artists?

No. The Art Recognition AI System can handle any artist, provided there are enough original artworks by that artist to train on.

Can the Art Recognition AI System distinguish between different styles of an artist?

Certainly. Our AI produces accurate results even for the artists which significantly changed their style across their careers, such as for example Cézanne. To be able to capture the different styles, we make sure to include a sufficient number of images from each period.

I believe my artwork belongs to a period which comprises less artworks than other periods by the assumed artist. Can the AI System handle this?

Indeed, the Art Recognition AI System can deal with this situation. The unique brushstroke of an artist spans across all styles, also on periods which contain less works. Therefore, if your artwork is an original piece by the assumed artist, the brushstroke will be recognized.

How large is the Art Recognition database?

Art Recognition has access to more than one million high resolution images.

What is the typical size of a dataset used for training?

Normally a dataset of around 200 images is enough to produce accurate results. For the more complex artists, this number should be higher in order to better capture the variations in style. We also augment the dataset by splitting the images into smaller patches, and zooming into them in order to capture the finer details.

Is the training data trustworthy? Where does Art Recognition get the data from?

We use images documented in Catalogue Raisonnés and embed the opinion of art experts whenever possible, in order to assemble training dataset of high quality.

How long does it take to receive a response?

Normally we deliver the AI Report and Certificate within 7 days. If we need longer, we inform you right away.

I have taken a photo of my artwork with my iPhone. Is the quality sufficient for the analysis?

Absolutely. Please make sure that the photo is upfront and with no shadows coming from lamps or other objects.