Skip to main content

Table 6 Final HoG descriptor size per B-scan, after building the image pyramid

From: Machine learning techniques for diabetic macular edema (DME) classification on SD-OCT images

Level of the pyramid

 

{1}

{2}

{3}

{4}

Feature vector size per B-Scan per level

266,112

63,468

15,120

3240

Total vector size per B-Scan

347,940