Very Deep Convolutional Networks for Large-Scale Image Recognition

4 Sep 2014Karen SimonyanAndrew Zisserman

In this work we investigate the effect of the convolutional network depth on its accuracy in the large-scale image recognition setting. Our main contribution is a thorough evaluation of networks of increasing depth using an architecture with very small (3x3) convolution filters, which shows that a significant improvement on the prior-art configurations can be achieved by pushing the depth to 16-19 weight layers... (read more)

PDF Abstract

Code


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Image Classification ImageNet VGG-19 Top 1 Accuracy 74.5% # 131
Top 5 Accuracy 92.0% # 90
Number of params 144M # 9
Image Classification ImageNet VGG-16 Top 1 Accuracy 74.4% # 132
Top 5 Accuracy 91.9% # 91
Number of params 138M # 10

Results from Other Papers


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK SOURCE PAPER COMPARE
Face Verification CK+ VGGFace Accuracy 92.20 # 5
Face Verification Oulu-CASIA VGGFace Accuracy 93.50 # 6
Activity Recognition In Videos DogCentric VGG [[Simonyan and Zisserman2015]] Accuracy 59.9 # 4
Image-to-Image Translation GTAV-to-Cityscapes Labels VGG16 60.3 mIoU 41.3 # 2

Methods used in the Paper