Breaking News English

Recently, I have no opportunities to use English. Especially speaking and listening are hardly used.

At such this time, the saviour who takes back my corrupt English proficiency.

www.breakingnewsenglish.com

A listening section is splits up into five stage. Slowest, Slower, British English, American English, Faster, Fastest.Many teaching materials of listening are distributed at present.But this site can’t only control a listening speed ,but also teach me various English.

It isn’t only that.Besides this also has the wonderful function.It’s called ‘speed reading’. Usually I’m not conscious of the speed which reads English.

When I read a specialized filed in English, the sentence are often difficult. Therefore I read a paper carefully.But that makes me to become slow in the speed which reads English.

I have to be conscious to increase the speed. This site would be useful at such time.

Gaussian

Gaussian Blur

写真にノイズ(小さな変則的な要素)が多くある場合、ガウスぼかしを使用し、画像の前処理 する。 [tex:{E=mc2}]

画像内の ピクセルを正方形(n2)のカーネルでたたみ込む ことで変換する。(全ピクセルに対し周囲のピクセルを考慮し平均値を出すことでなめらかなぼかしを表現する)

標準偏差σのガウシアンぼかしとは、n次元の入力画像A[i,j…]に対してn次元のガウス関数 [tex: \begin{math} G_\sigma(x,y,…)= \frac{1}{(\sqrt{2π\sigma2})}exp(-\frac{r2}{2\sigma2})(ここでr2=x2+y2+…) \end{math}]

の畳み込み和 [tex:( O(rn_cutoff) B[i,j,…]=\sum _{x,y,…\in Zn} G\sigma(x,y,…)A[i-x,j-y,…] )] をとることである

[tex:$ ガウス関数は中心から離れるにつれ値が小さくなり、r{cutoff} ~ 3\sigma程度で打つきり行列を作る。n次元ガウス関数はnこの一次元ガウス関数の直積によって表されることにより、ピクセルあたりの計算量を $] [tex:$ O(rn{cutoff}) $] から $ O(nr_{cutoff}) $ とぼかし処理を高速化することができる。

blurGaussian01.py

screen shot 2017-02-07 at 06 39 34

blurGaussian02.py

screen shot 2017-02-07 at 07 19 36