Skip to main content

Table 5 Scene classification rate (%) in MIT Indoor 67

From: DTCTH: a discriminative local pattern descriptor for image classification

Techniques

Accuracy

CNN-SVM, 2014 [92]

58.4

Places-CNN, 2014 [71]

68.24

ImageNet-CNN, 2014 [71]

56.79

Hybride-CNN, 2014 [71]

70.80

Dense SIFT (LSA + MMP) LSVM,

2011 [68]

44.19

dense SIFT (LLC + MP) LSVM,

2010 [62]

43.78

dense SIFT (LDC + LLC/LSA + MP)

LSVM, 2013 [64]

46.69

Dense SIFT (SSC + MP) OCL, 2012 [63]

44.35

Object Bank + LSVM, 2010 [55]

37.60

Dense SIFT (BoF) SVM with HI,

2014 [93]

45.86

DPM, 2011 [56]

30.40

CENTRIST (BoF) PmSVM-HI, 2012 [79]

47.15

CENTRIST (BoF) PmSVM- χ 2, 2012 [79]

46.20

PRICoLBP + SVM with χ 2, 2014 [8]

43.4

HOG, 2005 [56]

22.8

SPM, 2006 [8],

34.4

MM-scene, 2010 [94]

28.00

mCENTRIST (SPM) LSVM, 2014 [6]

44.6 ±1.2

mSIFT (SPM) LSVM, 2014 [6]

39.7 ±1.6

mGIST (SPM) LSVM, 2014 [6]

31.5 ±1.6

LGP (SPM) LSVM, 2013

34.24 ±1.12

OC-LBP (BoF) LSVM, 2013

36.99 ±2.34

LAID (SPM) LSVM, 2013

32.78 ±1.47

CLBP_S/M/C (SPM) LSVM, 2010

30.45 ±1.70

LTP (SPM) LSVM, 2010

35.87 ±1.23

GIST + LSVM, 2001

26.5 ±1.41

CENTRIST (SPM) LSVM, 2011

35.12 ±0.99

Proposed (DTCTH + LSVM)

43.33 ±0.72

Proposed (DTCTH + HI)

46.22 ±1.02