library(keras)
library(BiocManager)
library(imager)
## Loading required package: magrittr
##
## Attaching package: 'imager'
## The following object is masked from 'package:magrittr':
##
## add
## The following objects are masked from 'package:stats':
##
## convolve, spectrum
## The following object is masked from 'package:graphics':
##
## frame
## The following object is masked from 'package:base':
##
## save.image
library(EBImage)
##
## Attaching package: 'EBImage'
## The following objects are masked from 'package:imager':
##
## channel, dilate, display, erode, resize, watershed
library(tensorflow)
setwd('/Users/chipindulovemore/Documents')
pics<-c('cob_1.PNG','cob_2.PNG','cob_3.PNG','cob_4.PNG','cob_5.PNG','cob_6.PNG','cob_7.PNG','cob_8.PNG','cob_9.PNG','cob_10.PNG','cob_11.PNG','cob_12.PNG')
mypic<-list()
for(i in 1:12) {mypic[[i]]<-readImage(pics[i])}
print(mypic[[1]])
## Image
## colorMode : Color
## storage.mode : double
## dim : 209 629 4
## frames.total : 4
## frames.render: 1
##
## imageData(object)[1:5,1:6,1]
## [,1] [,2] [,3] [,4] [,5] [,6]
## [1,] 0.1568627 0.14117647 0.1960784 0.03921569 0.1215686 0.06274510
## [2,] 0.1960784 0.10980392 0.2078431 0.04313725 0.1803922 0.08235294
## [3,] 0.1372549 0.09019608 0.1725490 0.09411765 0.1960784 0.10980392
## [4,] 0.1607843 0.12941176 0.1137255 0.25490196 0.2235294 0.21176471
## [5,] 0.1333333 0.14509804 0.1019608 0.32156863 0.2156863 0.25490196
print(mypic[[8]])
## Image
## colorMode : Color
## storage.mode : double
## dim : 491 137 4
## frames.total : 4
## frames.render: 1
##
## imageData(object)[1:5,1:6,1]
## [,1] [,2] [,3] [,4] [,5] [,6]
## [1,] 0.8549020 0.8588235 0.8705882 0.8627451 0.8823529 0.8784314
## [2,] 0.8588235 0.8509804 0.8509804 0.8588235 0.8588235 0.8549020
## [3,] 0.8549020 0.8431373 0.8313725 0.8392157 0.8470588 0.8431373
## [4,] 0.8274510 0.8117647 0.8078431 0.8196078 0.8117647 0.8117647
## [5,] 0.8039216 0.8000000 0.8000000 0.7960784 0.8000000 0.8000000
hist(mypic[[8]])
display(mypic[[1]])
hist(mypic[[1]])
str(mypic)
## List of 12
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:209, 1:629, 1:4] 0.157 0.196 0.137 0.161 0.133 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:184, 1:473, 1:4] 0.0706 0.0627 0.0667 0.0784 0.098 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:212, 1:505, 1:4] 0.0392 0.0392 0.051 0.051 0.0627 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:181, 1:605, 1:4] 0.102 0.102 0.102 0.0863 0.0863 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:221, 1:696, 1:4] 0.0941 0.098 0.1255 0.1451 0.1686 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:197, 1:615, 1:4] 0.169 0.212 0.133 0.125 0.129 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:187, 1:523, 1:4] 0.0431 0.0275 0.0235 0.0235 0.0196 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:491, 1:137, 1:4] 0.855 0.859 0.855 0.827 0.804 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:408, 1:141, 1:4] 0.0745 0.0431 0.0431 0.1843 0.0745 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:520, 1:127, 1:4] 0.1059 0.0627 0.0471 0.051 0.0863 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:169, 1:620, 1:4] 0.0431 0.0588 0.0471 0.051 0.0471 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:215, 1:657, 1:4] 0.098 0.102 0.1098 0.0941 0.098 ...
## .. ..@ colormode: int 2
display(mypic[[1]])
print(mypic[[1]])
## Image
## colorMode : Color
## storage.mode : double
## dim : 209 629 4
## frames.total : 4
## frames.render: 1
##
## imageData(object)[1:5,1:6,1]
## [,1] [,2] [,3] [,4] [,5] [,6]
## [1,] 0.1568627 0.14117647 0.1960784 0.03921569 0.1215686 0.06274510
## [2,] 0.1960784 0.10980392 0.2078431 0.04313725 0.1803922 0.08235294
## [3,] 0.1372549 0.09019608 0.1725490 0.09411765 0.1960784 0.10980392
## [4,] 0.1607843 0.12941176 0.1137255 0.25490196 0.2235294 0.21176471
## [5,] 0.1333333 0.14509804 0.1019608 0.32156863 0.2156863 0.25490196
display(mypic[[3]])
summary(mypic[[1]])
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 0.0000 0.2706 0.6471 0.5949 1.0000 1.0000
hist((mypic[[1]]))
str(mypic)
## List of 12
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:209, 1:629, 1:4] 0.157 0.196 0.137 0.161 0.133 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:184, 1:473, 1:4] 0.0706 0.0627 0.0667 0.0784 0.098 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:212, 1:505, 1:4] 0.0392 0.0392 0.051 0.051 0.0627 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:181, 1:605, 1:4] 0.102 0.102 0.102 0.0863 0.0863 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:221, 1:696, 1:4] 0.0941 0.098 0.1255 0.1451 0.1686 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:197, 1:615, 1:4] 0.169 0.212 0.133 0.125 0.129 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:187, 1:523, 1:4] 0.0431 0.0275 0.0235 0.0235 0.0196 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:491, 1:137, 1:4] 0.855 0.859 0.855 0.827 0.804 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:408, 1:141, 1:4] 0.0745 0.0431 0.0431 0.1843 0.0745 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:520, 1:127, 1:4] 0.1059 0.0627 0.0471 0.051 0.0863 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:169, 1:620, 1:4] 0.0431 0.0588 0.0471 0.051 0.0471 ...
## .. ..@ colormode: int 2
## $ :Formal class 'Image' [package "EBImage"] with 2 slots
## .. ..@ .Data : num [1:215, 1:657, 1:4] 0.098 0.102 0.1098 0.0941 0.098 ...
## .. ..@ colormode: int 2
####Row Bind # converting the data set into one
trainx<-NULL
for(i in 1:5){trainx<-rbind(trainx,mypic[[i]])}## using the first five images of perfect maize cob (no abortion)
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 2)
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 2)
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 2)
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 2)
str(trainx) # stracture of train x (no abortion)
## num [1:5, 1:525844] 0.1569 0.0706 0.0392 0.102 0.0941 ...
## they are 12 pic , 6 no abortion and 6 abortions
## 5 pics will be used for training
## 6th pic will be used to test also the 12 th of abortions
for(i in 7:11){trainx<-rbind(trainx,mypic[[i]])}## using the first five images with abortion
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 2)
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 2)
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 2)
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 2)
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 2)
str(trainx) # stracture of train x (abortion)
## num [1:10, 1:525844] 0.1569 0.0706 0.0392 0.102 0.0941 ...
## test pic
testx<-rbind(mypic[[6]],mypic[[12]]) # 6th no abortion ,12th abortion
## Warning in rbind2(argl[[i]], r): number of columns of result is not a
## multiple of vector length (arg 1)
trainy<-c(0,0,0,0,0,1,1,1,1,1) # 0=no abortion and 1 = abortion
testy<-c(0,1) # 0= 6 and 1=12 that is no abortion and abortion respectively
## One hot Encoding
trainLabels<-to_categorical(trainy)
testLabels<-to_categorical(testy) # using tensorflow backened
trainLabels ### print
## [,1] [,2]
## [1,] 1 0
## [2,] 1 0
## [3,] 1 0
## [4,] 1 0
## [5,] 1 0
## [6,] 0 1
## [7,] 0 1
## [8,] 0 1
## [9,] 0 1
## [10,] 0 1
testLabels ### print
## [,1] [,2]
## [1,] 1 0
## [2,] 0 1
##Model
model<-keras_model_sequential()
model %>% ## the operator passes the command to the next line
layer_dense(units=256,activation ='relu',input_shape= c(525844)) %>%
layer_dense(units=128,activation='relu') %>%
layer_dense (units=2, activation = 'softmax')
summary (model)
## Model: "sequential"
## ___________________________________________________________________________
## Layer (type) Output Shape Param #
## ===========================================================================
## dense (Dense) (None, 256) 134616320
## ___________________________________________________________________________
## dense_1 (Dense) (None, 128) 32896
## ___________________________________________________________________________
## dense_2 (Dense) (None, 2) 258
## ===========================================================================
## Total params: 134,649,474
## Trainable params: 134,649,474
## Non-trainable params: 0
## ___________________________________________________________________________
#### compiling the model
model%>%
compile (loss='binary_crossentropy',
optimizer=optimizer_rmsprop(),
metrics=c('accuracy'))
# fit the model
history<-model%>%
fit(trainx,
trainLabels,
epochs=30,
batch_size=32,
validation_split=0.3)
plot(history)
####
## Evaluation and pediction-train-data
model %>% evaluate (trainx,trainLabels)
## $loss
## [1] 7.689546
##
## $acc
## [1] 0.5
pred<-model %>% predict_classes(trainx)
table(Predicted=pred,Actual=trainy) ### confusion matrix
## Actual
## Predicted 0 1
## 0 5 5
prob <-model%>% predict_proba(trainx) ### computing probabilities of belonging to a class
cbind(prob,Predicted=pred,Actual=trainy) ## combining probabilities
## Predicted Actual
## [1,] 1 0 0 0
## [2,] 1 0 0 0
## [3,] 1 0 0 0
## [4,] 1 0 0 0
## [5,] 1 0 0 0
## [6,] 1 0 0 1
## [7,] 1 0 0 1
## [8,] 1 0 0 1
## [9,] 1 0 0 1
## [10,] 1 0 0 1