In lovetoken/bsid: Background Subtraction and Image Denoising (BSID)

library(bsid)
library(useful.lovetoken)
pacman::p_load(dplyr, bmp, pixmap, animation, corpcor, progress, servr, knitr)
opts_chunk$set(fig.path='figure/', warning = F, message = F)

creating M matrix

M matrix 만들기 creat_vm()

creat_vm(path="./example_data/airport_sub/")
M <- bsid_env$M

$M$ 확인하기 save_anipic()

save_anipic(M, save.name="airport_sub_M")

PCP algorithm

Principal Component Pursuit algorithm 를 이용하여 L, S matrix 만들기 PCP()

PCP(M)
L <- bsid_env$L
S <- bsid_env$S

$L, S$ 확인하기

save_anipic(L, save.name="airport_sub_L")
save_anipic(S, save.name="airport_sub_S")

$L$ matrix

$S$ matrix

M matrix 에 PWN_{Point type White Noise} 추가하기

임의위치의 픽셀값을 1로 부여하여 잡음을 추가 WN_point()

WN_point(M)
Mstar <- bsid_env$Mstar

$M^\star$ 확인하기

save_anipic(Mstar, save.name="airport_sub_Mstar")

Mstar PCP

PCP(Mstar)
Lstar <- bsid_env$L
Sstar <- bsid_env$S

save_anipic(Lstar, save.name="airport_sub_Lstar")
save_anipic(Sstar, save.name="airport_sub_Sstar")

$L^\star$ 확인하기

$S^\star$ 확인하기

Image Denoising

$M^\prime$ 만들기 (method="MDMR_filter")

denoising(Mstar, method="MDMR_filter", W0=1, lambda=0.1)
Mprime <- bsid_env$prime

save_anipic(Mprime, save.name="airport_sub_Mprime")

$M^\prime$ 확인하기

$S^\prime$ 만들기 (method="median_filter")

denoising(Sstar, method="MDMR_filter", W0=1, lambda=0.1)
Sprime <- bsid_env$prime

save_anipic(Sprime, save.name="airport_sub_Sprime")

$S^\prime$ 확인하기

$S^\prime + L^\star$ 확인하기

Mpprime <- Sprime + Lstar

save_anipic(Mpprime, save.name="airport_sub_Mpprime")

MSE compare

$$\mid\mid{M-M^{\prime}}\mid\mid_{F}$$

norm(M-Mprime, type="F")

$$\mid\mid{M-(S^{\prime} + L^{\star})}\mid\mid_{F}$$

norm(M-Mpprime, type="F")

$$\mid\mid{M-M^{\prime}}\mid\mid_{F} \ \ge\ \mid\mid{M-(S^{\prime} + L^{\star})}\mid\mid_{F}$$

• email : [e-mail]
• blog : https://lovetoken.github.io/

lovetoken/bsid documentation built on May 21, 2019, 7:49 a.m.