The Computer Vision’s Battleground: Choose Your Champion

Author(s): Salvatore Raieli

Originally published on Towards AI.

Which is the best computer vision model? Which one is best for a particular task?
Photo by GR Stocks on Unsplash

Transfer learning has changed computer vision, but many open questions remain. For example, what is the best architecture? Which one is best for a task? Every article claims to be state-of-the-art, but really? Here, a study has determined this empirically, an answer to the practical questions every artificial intelligence practitioner asks.

Photo by Mika Matin on Unsplash

The dominant paradigm in computer vision states that a system consists of a backbone (a feature extractor network) followed by a head that is task-specific. The backbone can produce either an array of features for object detection and localization or… Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI