Started thinking about padding

Author: xd009642

src/processing/conv.rs

@@ -6,6 +6,23 @@ use num_traits::{Num, NumAssignOps};
 use std::marker::PhantomData;
 use std::marker::Sized;
 
+pub struct NoPadding;
+pub struct ZeroPadding;
+pub struct ExtendPadding;
+
+pub trait ImagePadder where Self: Sized + Copy {
+    type Data;
+
+    fn pad(&self, pad_size: (usize, usize)) -> Self;
+}
+
+
+impl<T> ImagePadder<T> for NoPadding {
+    fn pad(&self, pad_size: (usize, usize)) -> Self {
+        *self
+    }
+}
+
 /// Perform image convolutions
 pub trait ConvolutionExt
 where