Se ha producido un error
Tenemos problemas para reproducir este vídeo.

AI for Commerce

Find out how AI powers new shopping innovations and makes life easier for both shoppers and sellers.

Advancing AI to make shopping easier for everyone

Product recognition is among the most important ways to make it easier for people to shop online today. Learn how Facebook AI is building the world's largest shoppable social media platform, where billions of items can be bought and sold in one place.
Se ha producido un error
Tenemos problemas para reproducir este vídeo.

Computer vision through the years at Facebook

2013

Facebook AI lab opens

Identifying “person & dog”

Se ha producido un error
Tenemos problemas para reproducir este vídeo.

2016

SharpMask

Segmenting objects at pixel-level

Se ha producido un error
Tenemos problemas para reproducir este vídeo.

2016

Feature Pyramid Networks

Processing large and small objects efficiently

Se ha producido un error
Tenemos problemas para reproducir este vídeo.

2017

Mask R-CNN

Achieving state-of-the-art object segmentation and classification

Se ha producido un error
Tenemos problemas para reproducir este vídeo.

2019

Panoptic FPN

Recognizing background and foreground simultaneously

Se ha producido un error
Tenemos problemas para reproducir este vídeo.

2019

Mesh R-CNN

Achieving state-of-the-art in 3D object recognition

Se ha producido un error
Tenemos problemas para reproducir este vídeo.

2019

Instance Mask Projection

Identify clothing behind layers

Se ha producido un error
Tenemos problemas para reproducir este vídeo.

2020

GrokNet

Building a universal product recognition model

Se ha producido un error
Tenemos problemas para reproducir este vídeo.

2021

Multimodal Product Understanding

Incorporating signals from associated text

Read the Latest Publication

RESEARCH

COMPUTER VISION

Large-Scale Weakly-Supervised Attribute-Object Composition Learning

In this paper, we study learning to predict the combination of object and attribute class labels, expanding the number of possible classes to the order of hundreds of thousands or more. The learning is performed on large-scale image dataset with noisy attribute-object labels.

Filip Radenovic, Animesh Sinha, Albert Gordo, Tamara Berg, Dhruv Mahajan

Help Us Pioneer the Future of AI

We share our open source frameworks, tools, libraries, and models for everything from research exploration to large-scale production deployment.