Pratyush Mahapatra

MediaAttention-Ads v1

Where attention may concentrate in an ad

Open in Studio

What to provide

Attach a static PNG, JPEG or WebP image up to 3 MB and 12 megapixels.

How it works

UNISAL fine-tuned on UEyes estimates visual saliency. Studio overlays the map and measures the share of estimated attention in each of the three horizontal sections.

How to read the result

Compare two versions of a poster to see which elements compete for attention. The colors and percentages describe the model output. Measure actual gaze, clicks and conversions with your own observations.

Credits and licenses

MediaAttention-Ads v1Apache 2.0

Model by Pratyush Mahapatra, based on UNISAL (Droste, Jiao and Noble, ECCV 2020) and fine-tuned on UEyes.

UEyes: Understanding Visual Saliency across User Interface Types by Yue Jiang, Luis A. Leiva, Hamed Rezazadegan Tavakoli, Paul R. B. Houssel, Julia Kylmälä and Antti Oulasvirta (CHI 2023). Both attention models were fine-tuned on this dataset, released under CC BY 4.0.

Official documentation