Table 3.

Summary of publications focused on feature selection of prostate histopathology images.

Features Type	Reference	Year	Accuracy Result
Texture	[56]	2011	The AUC value is 0.91 for the first database and 0.96 for the second database.
	[102]	2015	The proposed method outperforms the classic SVM-RFE in accuracy and reducing redundancy.
	[103]	2018	The proposed method attained a classification accuracy around 99%.
Topological	[13]	2011	The model attainted an average accuracy 90%.
	[50]	2011	The test classification results have an average of 96.76%
	[49]	2017	The developed way achieved 93.0% training accuracy and 97.6% testing accuracy, for the tested cases.
Morphological	[15]	2007	Average accuracy for prostate cancer classification was 92.48%
	[104]	2011	The system achieved 0.55 under the precision recall curve measure
	[58]	2019	The prediction model resulted an average accuracy of 90.2%
Color	[98]	2012	The proposed method attained an average of 86% accuracy in classifying a tissue pattern into different classes.
	[105]	2006	They achieved accuracy of 91.3%
Color & Texture	[106]	2012	The algorithm achieved an average of 86% and 93% of classification accuracy.
	[107]	2012	Classification accuracies are 97.6%, 96.6% and 87.3% when differentiating Gleason 4 versus Gleason 3, Gleason 5 versus Gleason 3, and Gleason 5 versus Gleason 4.
Topological & Morphological & Texture	[48]	2007	SVM classifier applied to test the accuracy of the extracted features and achieved about 93% when differentiating among Gleason grade 3 and stroma, 92.4% among epithelium and stroma, and 76.9% among Gleason 4 and 3.
	[27]	2019	The proposed model using hand-crafted features achieved an average accuracy of 94.6%.