Large-Scale Coarse-to-Fine Object Retrieval Ontology and Deep Local Multitask Learning

Algorithm 6

Conversion of the score vector to the binary vector.

	Procedure: BINARY_CONVERTER
	Input:
	testPred //an array of multilabel predictions of each sample in the test set
	bestThresholds //an array of best thresholds of each label
	Output:
	binaryTestPred //an array of multilabel predictions of each sample.
	Goal:
	Convert the score vector to the binary vector in an MCC imbalanced data problem solver
	BEGIN
	for i = 0 to numberOfSample do:
	begin
	[binaryConverted] ⟵ [NULL]
	#binary converter for each prediction
	for j = 0 to numberOfLabel do:
	begin
	if testPred [i][j] >= bestThresholds [j]:
	binaryConverted ⟵ 1
	else:
	binaryConverted ⟵ 0
	end
	binaryTestPred ⟵ binaryConverted   #show overall prediction
	end
	END