Add filter_2d

src/third_party/image/kernel.rs

@@ -0,0 +1,54 @@
+use crate::base::DMatrix;
+use crate::storage::Storage;
+use crate::{Dim, Dynamic, Matrix, Scalar};
+use num::Zero;
+use std::ops::{AddAssign, Mul};
+
+impl<N, R1, C1, SA> Matrix<N, R1, C1, SA>
+where
+    N: Scalar + Zero + AddAssign + Mul<Output = N> + Copy,
+    R1: Dim,
+    C1: Dim,
+    SA: Storage<N, R1, C1>,
+{
+    /// Returns the convolution of the target matrix and a kernel.
+    ///
+    /// # Arguments
+    ///
+    /// * `kernel` - A Matrix with size > 0
+    ///
+    /// # Errors
+    /// Inputs must satisfy `matrix.len() >= matrix.len() > 0`.
+    ///
+    pub fn filter_2d<R2, C2, SB>(&self, kernel: Matrix<N, R2, C2, SB>) -> DMatrix<N>
+    where
+        R2: Dim,
+        C2: Dim,
+        SB: Storage<N, R2, C2>,
+    {
+        let mat_shape = self.shape();
+        let ker_shape = kernel.shape();
+
+        if ker_shape == (0, 0) || ker_shape > mat_shape {
+            panic!("filter_2d expects `self.shape() >= kernel.shape() > 0`, received {:?} and {:?} respectively.", mat_shape, ker_shape);
+        }
+
+        let result_shape = (mat_shape.0 - ker_shape.0 + 1, mat_shape.1 - ker_shape.1 + 1);
+        let mut conv = DMatrix::zeros_generic(
+            Dynamic::from_usize(result_shape.0),
+            Dynamic::from_usize(result_shape.1),
+        );
+
+        // TODO: optimize
+        for i in 0..(result_shape.0) {
+            for j in 0..(result_shape.1) {
+                for k in 0..(ker_shape.0) {
+                    for l in 0..(ker_shape.1) {
+                        conv[(i, j)] += self[(i + k, j + l)] * kernel[(k, l)]
+                    }
+                }
+            }
+        }
+        conv
+    }
+}

src/third_party/image/mod.rs

@@ -0,0 +1 @@
+mod kernel;

src/third_party/mod.rs

@@ -4,3 +4,5 @@ mod alga;
 mod glam;
 #[cfg(feature = "mint")]
 mod mint;
+
+mod image;

tests/lib.rs

@@ -22,6 +22,7 @@ mod linalg;
 
 #[cfg(feature = "proptest-support")]
 mod proptest;
+mod third_party;
 
 //#[cfg(feature = "sparse")]
 //mod sparse;

tests/third_party/image/kernel.rs

@@ -0,0 +1,16 @@
+use na::{Matrix3, MatrixMN, U10, U8};
+use std::panic;
+
+#[test]
+fn image_convolve_check() {
+    // Static Tests
+    let expect = MatrixMN::<usize, U8, U8>::from_element(18);
+    let src = MatrixMN::<usize, U10, U10>::from_element(2);
+    let kernel = Matrix3::<usize>::from_element(1);
+    let result = src.filter_2d(kernel);
+    println!("src: {}", src);
+    println!("ker: {}", kernel);
+    println!("res: {}", result);
+
+    assert_eq!(result, expect);
+}

tests/third_party/image/mod.rs

@@ -0,0 +1 @@
+mod kernel;

tests/third_party/mod.rs

@@ -0,0 +1 @@
+mod image;