Large Scale Visual Recognition Challenge 2014 results announced

[madeye]

最近几年 Deep Learning 大行其道，几乎每年 ILSVRC 的结果都有大幅提升。于是今年 Object detection 的精度从 22.6 % 提高到了43.9%，Classification error 从 11.2% 降到了6.7%。目测两到三年内各种激动人心的实际应用也会大爆发。

From: ILSVRC 2014 [mailto:[e-mail]]
Sent: Monday, August 18, 2014 8:48 PM
To: ILSVRC 2014
Subject: ILSVRC2014 results announced

Dear ILSVRC2014 participants,

It is our great pleasure to release to you the results of the ILSVRC2014 challenge:

http://image-net.org/challenges/LSVRC/2014/results

We had an unprecedented number of participants this year: 38 teams with 126 entries -- more than 50% increase from last year! Object detection average precision is up to 43.9% from 22.6% last year. Classification error is down to 6.7% compared to 11.2% last year. The challenge is even featured in the New York Times!

We invite you all to come hear about these approaches at the ILSVRC workshop on September 12th at ECCV2014 (details coming soon):

http://image-net.org/challenges/LSVRC/2014/eccv2014

This year we’re going to be giving all teams that submitted to ILSVRC2014 an opportunity to participate in the workshop. We will be in touch soon; please look out for an email from us.

Thank you for your participation in ILSVRC2014, and hope to see you in Zurich!

Best,

ILSVRC2014 organizers