Visual Recognition Software for Binary Classification and Its Application to Spruce Pollen Identification
Fig 4
An image of a black spruce grain is shown with the following color quantizations to reduce the number of possible pixel intensity values: (A) 128x, (B) 64x, (C) 32x, (D) 16x, (E) 8x, and (F) 1x. A color quantization of 128x reduces the possible intensity values to 0 and 1. A color quantization of 1x is the original image. The pollen grain diameter is ~80 μm.