Classification of SD-OCT Volumes Using Local Binary Patterns: Experimental Validation for DME Detection

Table 4

The outline and summary of the performed experiments.

Common	Dataset	Preprocessing	Features	Mapping	Representation	Classification	Evaluation
	SERI	NLM	LBP, LBP-TOP				Leave-One-Patient Out Cross-Validation (LOPO-CV); SE, SP
Baseline [12]; Goal: evaluation of features, mapping, and representation	+ Duke	~	~	Global; Local	BoW; Histogram	RF	+ comparison with [10]
Experiment; Goal: finding the optimum number of words	~	+; ++	~	Global; Local	BoW	LR	+ACC, F1 score (F1)
Experiment; Goal: evaluation of different preprocessing and classifiers for high-level features	~	+; ++	~	Global; Local	BoW; Optimal	3-NN; RF; SVM; GB	~
Experiment; Goal: evaluation of different preprocessing and classifiers for low-level features	~	+; ++	~	Global	Histogram	3-NN; RF; SVM; GB	~

 ~ indicates that common configuration applies.