FilesExpand file tree

 master

/

testingScript.R

Copy path

Blame

More file actions

Blame

More file actions

Latest commit

 

History

History

History

203 lines (183 loc) · 5.91 KB

 master

/

testingScript.R

Copy path

Top

File metadata and controls

• Code
• Blame

203 lines (183 loc) · 5.91 KB

Raw

Copy raw file

Download raw file

Open symbols panel

Edit and raw actions

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

42

43

44

45

46

47

48

49

50

51

52

53

54

55

56

57

58

59

60

61

62

63

64

65

66

67

68

69

70

71

72

73

74

75

76

77

78

79

80

81

82

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

99

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

# Load Data

library(readr)

library(bmp)

library(pixmap)

library(dplyr)

#install.packages("readbitmap")

#library(readbitmap)

# location of train and test images

train_files = list.files(path="E:\\data\\train\\train\\.", pattern=".Bmp",all.files=T, full.names=T, no.. = T)

test_files = list.files(path="E:\\data\\test\\test\\.", pattern=".Bmp",all.files=T, full.names=T, no.. = T)

# we will create a matrix with a row for each image.

# since images are 20 by 20 pixels, our new matrix will have 400 columns of pixel data.

train_images = list(rep(0,length(train_files)))

#train_images

train_mat = matrix(ncol = 401, nrow = length(train_files))

#write.csv(train_mat)

#write.csv(train_mat, file = "E:\\data\\MyTrainMatInitData.csv")

#length(train_files)

#View(as.data.frame(train_files[[1]]))

#train_mat

#View(train_mat)

stp <- strsplit(train_files[1], split = ".", fixed=T)

#

# stp

#

ulist <- unlist(stp)[2]

# ulist

r = as.data.frame(img[ , , 1])

g = as.data.frame(img[ , , 2])

b = as.data.frame(img[ , , 3])

temp = (r + g + b)/3

#temp

# temp = r

temp = as.vector(temp)

temp = c(ulist, temp)

temp = as.vector(temp)

#write.csv(temp, file = "E:\\data\\MyTrainTempInit.csv")

View(temp)

summary(temp$band1)

for(i in 2:401)

train_mat[1,i] = temp$band1[i]

View(temp$band1)

View(train_mat)

count <- 0

count1 <- 0

count2 <- 0

count3 <- 0

for(i in 1:nrow(train_mat)) {

ID = unlist(strsplit(train_files[i], split = ".", fixed=T))[2]

#ID2 = as.numeric(unlist(strsplit(ID, split = "/"))[2])

ID2 = as.vector(unlist(strsplit(ID, split = "/"))[2])

train_images[[i]]= readGDAL(train_files[i])

img = train_images[[i]]

if (length(dim(img)) == 3) {

# r = img[ , , 1]

# g = img[ , , 2]

# b = img[ , , 3]

# temp = (r + g + b)/3

# temp = as.vector(temp)

# temp = c(ID2, temp)

# train_mat[i,] = temp

count3 <- count3 + 1

}

else if (length(dim(img)) == 2) {

count <- count + 1

r = as.data.frame(img[ , , 1])

g = as.data.frame(img[ , , 2])

b = as.data.frame(img[ , , 3])

temp = (r + g + b)/3

# temp = r

temp = as.vector(temp)

temp = c(ID2, temp)

for(j in 2:401){

train_mat[i,j] = temp$band1[j]

}

#temp = as.vector(img[ , , ])

#temp = c(ID2, temp)

#train_mat[i,] = temp

}

else if (length(dim(img)) == 1) {

count1 <- count1 + 1

#temp = as.vector(img[ , , ])

#temp = c(ID2, temp)

#train_mat[i,] = temp

} else if (length(dim(img)) == 0) {

count2 <- count2 + 1

#temp = as.vector(img[ , , ])

#temp = c(ID2, temp)

#train_mat[i,] = temp

}

}

class(ID2)

count

count1

count2

count3

# check dimensions of matrix and to make sure everything loaded properly (no NAs)

dim(train_mat)

c <- sum(is.na(train_mat))

c

count <- 0

count1 <- 0

count2 <- 0

count3 <- 0

# and for the test images also

test_images = list(rep(0,length(test_files)))

test_mat = matrix(ncol = 401, nrow = length(test_files))

for(i in 1:nrow(test_mat)) {

ID = unlist(strsplit(test_files[i], split = ".", fixed=T))[2]

ID2 = as.numeric(unlist(strsplit(ID, split = "/"))[2])

test_images[[i]] = readGDAL(test_files[i])

img = test_images[[i]]

if (length(dim(img)) == 3) {

# r = as.data.frame(img[ , , 1])

# g = as.data.frame(img[ , , 2])

# b = as.data.frame(img[ , , 3])

# temp = (r + g + b)/3

# temp = as.vector(temp)

# temp = c(ID2, temp)

# test_mat[i,] = temp

count3 <- count3 + 1

}

else if (length(dim(img)) == 2) {

count <- count + 1

r = as.data.frame(img[ , , 1])

g = as.data.frame(img[ , , 2])

b = as.data.frame(img[ , , 3])

temp = (r + g + b)/3

temp = as.vector(temp)

temp = c(ID2, temp)

test_mat[i,] = temp

#temp = as.vector(img[ , , ])

#temp = c(ID2, temp)

#train_mat[i,] = temp

}

else if (length(dim(img)) == 1) {

count1 <- count1 + 1

#temp = as.vector(img[ , , ])

#temp = c(ID2, temp)

#train_mat[i,] = temp

} else if (length(dim(img)) == 0) {

count2 <- count2 + 1

#temp = as.vector(img[ , , ])

#temp = c(ID2, temp)

#train_mat[i,] = temp

}

}

count

count1

count2

# check dimensions of matrix and to make sure everything loaded properly (no NAs)

dim(test_mat)

d <- sum(is.na(test_mat))

# convert matrices do data frames and delete other matrices, vectors, and lists from memory

trainDF = as.data.frame(train_mat)

trainDF = trainDF %>% dplyr::rename(ID = V1)

testDF = as.data.frame(test_mat)

testDF = testDF %>% dplyr::rename(ID = V1)

rm(r, g, b, temp, i, img, ID, ID2, train_mat, test_mat, train_images, test_images)

# Read in the training labels and merge with the trainDF

trainLabels = read_csv("E:\\data\\trainLabels.csv", col_types = cols(ID = col_character()))

trainLabels$ID = as.numeric(trainLabels$ID)

trainDF = merge(x = trainDF, y = trainLabels, by = "ID") # merge by common ID columns

trainDF = trainDF[, c(1,402,2:401)] # rearrange columns in dataframe so that ID is first, Class is second, ...

trainDF$Class = as.factor(trainDF$Class)

# Duplicates - change indices to 3:402 since I just loaded the data

dup1 = which(duplicated(trainDF[,3:402]))

dup2 = which(duplicated(trainDF[,3:402], fromLast = TRUE))

dup_rows = sort(c(dup1, dup2))

trainDF[dup_rows,1:5]

# row 3055 is classified as a "1", but it has the same pixel values as row 4106, which is classified as an "E"

# it's definitely a 1, so we will remove row 4106

trainDF = trainDF[-c(2297, 3226, 4106, 4598, 4820, 5208, 6270), ]

# display image of specified row of trainDF

row_num = 4165

m = matrix(unlist(trainDF[row_num,-c(1,2)]),nrow = 20,byrow = F)

image(m,col=grey.colors(255), xlab = trainDF$Class[row_num])

# everything checks out, so I will remove the ID column

trainDF = trainDF[,-1]

rm(trainLabels)

# write trainDF and testDF to R Data Sets

saveRDS(testDF, "E:\\data\\testDF2.RDS")

saveRDS(trainDF, "E:\\data\\trainDF2.RDS")