Multi-column deep neural network beats humans with 99.46% at IJCNN 2011 GTSRB

At the IJCNN 2011 GTSRB traffic sign recognition competition, a multi-column deep neural network scored 99.46%, above the human average of 98.84%

Why it matters

An officially recognized benchmark record of a neural network surpassing humans on a specific visual recognition task

Source — Primary sources first; secondary reports are labelled “Press”

Officialbenchmark.ini.rub.dehttps://benchmark.ini.rub.de/gtsrb_results.html

Added in Week 35, 2026 · See that week

Nearby on the same axis

Building high-level features using large scale unsupervised learningPaperDeep Neural Networks for Acoustic Modeling in Speech RecognitionOfficialRepresentation Learning: A Review and New PerspectivesPaperImproving neural networks by preventing co-adaptation of feature detectorsPaper