diff --git a/example/hessian.cpp b/example/hessian.cpp
index e27199f5f4..cf05050bcc 100644
--- a/example/hessian.cpp
+++ b/example/hessian.cpp
@@ -60,8 +60,8 @@ gil::gray8_image_t to_grayscale(gil::rgb8_view_t original)
 
 void apply_gaussian_blur(gil::gray8_view_t input_view, gil::gray8_view_t output_view)
 {
-    constexpr static auto filter_height = 5ull;
-    constexpr static auto filter_width = 5ull;
+    constexpr static std::ptrdiff_t filter_height = 5ull;
+    constexpr static std::ptrdiff_t filter_width = 5ull;
     constexpr static double filter[filter_height][filter_width] =
     {
         { 2,  4,  6,  4,  2 },
diff --git a/example/mandelbrot.cpp b/example/mandelbrot.cpp
index 0bcfc2ea14..1aa5a90897 100644
--- a/example/mandelbrot.cpp
+++ b/example/mandelbrot.cpp
@@ -40,7 +40,7 @@ struct mandelbrot_fn
         t=pow(t,0.2);
 
         value_type ret;
-        for (int k=0; k<num_channels<P>::value; ++k)
+        for (std::size_t k=0; k<num_channels<P>::value; ++k)
             ret[k]=(typename channel_type<P>::type)(_in_color[k]*t + _out_color[k]*(1-t));
         return ret;
     }
diff --git a/include/boost/gil/channel_algorithm.hpp b/include/boost/gil/channel_algorithm.hpp
index 12ecd01b0a..848c9652ad 100644
--- a/include/boost/gil/channel_algorithm.hpp
+++ b/include/boost/gil/channel_algorithm.hpp
@@ -79,7 +79,7 @@ struct unsigned_integral_max_value<packed_channel_value<K>>
 
 template <typename UnsignedIntegralChannel>
 struct unsigned_integral_num_bits
-    : std::integral_constant<int, sizeof(UnsignedIntegralChannel) * 8>
+    : std::integral_constant<int, static_cast<int>(sizeof(UnsignedIntegralChannel) * 8)>
 {};
 
 template <int K>
diff --git a/include/boost/gil/color_base.hpp b/include/boost/gil/color_base.hpp
index a05b409de8..b638aacbba 100644
--- a/include/boost/gil/color_base.hpp
+++ b/include/boost/gil/color_base.hpp
@@ -84,7 +84,7 @@ struct homogeneous_color_base<Element, Layout, 1>
         typename U = Element,
         typename = typename std::enable_if<!std::is_reference<U>::value>::type
     >
-    homogeneous_color_base() : v0_{} {};
+    homogeneous_color_base() : v0_{} {}
 
     explicit homogeneous_color_base(Element v) : v0_(v) {}
 
diff --git a/include/boost/gil/color_convert.hpp b/include/boost/gil/color_convert.hpp
index 080d758ae9..a4b6427023 100644
--- a/include/boost/gil/color_convert.hpp
+++ b/include/boost/gil/color_convert.hpp
@@ -170,13 +170,13 @@ struct default_color_converter_impl<rgb_t, cmyk_t>
         // Apply color correction, strengthening, reducing non-zero components by
         // s = 1 / (1 - k) for k < 1, where 1 denotes dst_t max, otherwise s = 1 (literal).
         uint_t const dst_max = channel_traits<uint_t>::max_value();
-        uint_t const s_div   = dst_max - k;
+        uint_t const s_div   = static_cast<uint_t>(dst_max - k);
         if (s_div != 0)
         {
             double const s = dst_max / static_cast<double>(s_div);
-            c = (c - k) * s;
-            m = (m - k) * s;
-            y = (y - k) * s;
+            c = static_cast<uint_t>((c - k) * s);
+            m = static_cast<uint_t>((m - k) * s);
+            y = static_cast<uint_t>((y - k) * s);
         }
         else
         {
diff --git a/include/boost/gil/image_processing/harris.hpp b/include/boost/gil/image_processing/harris.hpp
index c5deedda99..18185afd81 100644
--- a/include/boost/gil/image_processing/harris.hpp
+++ b/include/boost/gil/image_processing/harris.hpp
@@ -44,7 +44,7 @@ void compute_harris_responses(
             " tensor from the same image");
     }
 
-    auto const window_length = weights.size();
+    std::ptrdiff_t const window_length = weights.size();
     auto const width = m11.width();
     auto const height = m11.height();
     auto const half_length = window_length / 2;
diff --git a/include/boost/gil/image_processing/hessian.hpp b/include/boost/gil/image_processing/hessian.hpp
index 2c8c3adeb1..f88c153607 100644
--- a/include/boost/gil/image_processing/hessian.hpp
+++ b/include/boost/gil/image_processing/hessian.hpp
@@ -50,9 +50,9 @@ inline void compute_hessian_responses(
             auto ddxx_i = channel_t();
             auto ddyy_i = channel_t();
             auto dxdy_i = channel_t();
-            for (typename OutputView::coord_t w_y = 0; w_y < weights.size(); ++w_y)
+            for (typename OutputView::coord_t w_y = 0; w_y < static_cast<std::ptrdiff_t>(weights.size()); ++w_y)
             {
-                for (typename OutputView::coord_t w_x = 0; w_x < weights.size(); ++w_x)
+                for (typename OutputView::coord_t w_x = 0; w_x < static_cast<std::ptrdiff_t>(weights.size()); ++w_x)
                 {
                     ddxx_i += ddxx(x + w_x - center, y + w_y - center)
                         .at(std::integral_constant<int, 0>{}) * weights.at(w_x, w_y);
diff --git a/include/boost/gil/image_processing/numeric.hpp b/include/boost/gil/image_processing/numeric.hpp
index 2109c555f4..b601b17dd0 100644
--- a/include/boost/gil/image_processing/numeric.hpp
+++ b/include/boost/gil/image_processing/numeric.hpp
@@ -48,7 +48,7 @@ inline double lanczos(double x, std::ptrdiff_t a)
     if (0 <= x && x <= 0)
         return 1;
 
-    if (-a < x && x < a)
+    if (static_cast<double>(-a) < x && x < static_cast<double>(a))
         return normalized_sinc(x) / normalized_sinc(x / static_cast<double>(a));
 
     return 0;