# Very Deep Convolutional Networks for Large-Scale Image Recognition

**Authors:** Karen Simonyan, Andrew Zisserman  
**arXiv:** [1409.1556](https://arxiv.org/abs/1409.1556)  
**Date:** 2014-09-04  
**Summary page:** https://www.summarizepaper.com/en/paper/1409.1556v1/

## Key points

- Authors Karen Simonyan and Andrew Zisserman studied the impact of convolutional network depth on accuracy in large-scale image recognition.
- Increasing the number of weight layers to 16-19 resulted in significant improvements compared to previous configurations.
- Their research helped them secure first place in localization and second place in classification tracks at the ImageNet Challenge 2014.
- Network depth plays a crucial role in improving image recognition performance.
- This study provides valuable insights for future advancements in the field.

## Summary

In their work titled "Very Deep Convolutional Networks for Large-Scale Image Recognition," authors Karen Simonyan and Andrew Zisserman investigated the impact of convolutional network depth on accuracy in the field of large-scale image recognition. Through a comprehensive evaluation of networks with increasing depth, they found that by increasing the number of weight layers to 16-19, significant improvements can be achieved compared to prior-art configurations. These findings were instrumental in their submission to the ImageNet Challenge 2014 where their team secured first place in localization and second place in classification tracks. This research highlights the importance of network depth in improving image recognition performance and provides valuable insights for future advancements in this field.