iT邦幫忙

第 12 屆 iThome 鐵人賽

DAY 26
1
AI & Data

30天只學U-net系列 第 26

[day-26] U-net Experiments (2) - performance

  • 分享至 

  • xImage
  •  

前言

說明了Unet的performance。

Performance

The u-net (averaged over 7 rotated versions of the input data) achieves without any further pre- or postprocessing a warping error of 0.0003529 (the newbest score, see Table 1) and a rand-error of 0.0382.

僅需要利用旋轉(8倍樣本)就可以獲得誤差。(參考表格)
https://ithelp.ithome.com.tw/upload/images/20201010/20112571o6st8VKfjR.png

基本上Unet 在 Warping 是最好的,上面 human values 代表人類的表現,目前沒有超越人類的跡象(在這個領域)。

This is significantly better than the sliding-window convolutional network result by Ciresan et al. [1], whose best submission had a warping error of 0.000420 and a rand error of 0.0504. In terms of rand error the only better performing algorithms on this data set use highly data set specific post-processing methods1 applied to the probability map of Ciresan et al. [1].

比文獻1的成果好很多。

Concusion

闡述了成果有多好,但是沒有想像中的那麼好呢!!

Reference

[0] U-net
[1] : Deep neural networks segment neuronal membranes in electron microscopy images.


上一篇
[day-25] U-net Experiments (1) - rule
下一篇
[day-27] U-net Experiments (3) - performance 2
系列文
30天只學U-net30
圖片
  直播研討會
圖片
{{ item.channelVendor }} {{ item.webinarstarted }} |
{{ formatDate(item.duration) }}
直播中

尚未有邦友留言

立即登入留言