diff --git a/src/core/CL/CLKernelLibrary.cpp b/src/core/CL/CLKernelLibrary.cpp
index 75ff248..ef3a431 100644
--- a/src/core/CL/CLKernelLibrary.cpp
+++ b/src/core/CL/CLKernelLibrary.cpp
@@ -364,6 +364,8 @@
     { "softmax_layer_max_shift_exp_sum_quantized_serial", "softmax_layer_quantized.cl" },
     { "softmax_layer_max_shift_exp_sum_quantized_parallel", "softmax_layer_quantized.cl" },
     { "softmax_layer_max_shift_exp_sum_serial", "softmax_layer.cl" },
+    { "space_to_batch", "space_to_batch.cl" },
+    { "space_to_batch_static", "space_to_batch.cl" },
     { "softmax_layer_max_shift_exp_sum_parallel", "softmax_layer.cl" },
     { "strided_slice", "slice_ops.cl" },
     { "suppress_non_maximum", "canny.cl" },
@@ -760,6 +762,10 @@
 #include "./cl_kernels/slice_ops.clembed"
     },
     {
+        "space_to_batch.cl",
+#include "./cl_kernels/space_to_batch.clembed"
+    },
+    {
         "tablelookup.cl",
 #include "./cl_kernels/tablelookup.clembed"
     },
diff --git a/src/core/CL/cl_kernels/space_to_batch.cl b/src/core/CL/cl_kernels/space_to_batch.cl
new file mode 100644
index 0000000..1343695
--- /dev/null
+++ b/src/core/CL/cl_kernels/space_to_batch.cl
@@ -0,0 +1,151 @@
+/*
+ * Copyright (c) 2018 ARM Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software withoutput restriction, including withoutput limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KOUTD, EXPRESS OR
+ * IMPLIED, OUTCLUDOUTG BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONOUTFROUTGEMENT. OUT NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER OUT AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISOUTG FROM,
+ * OUT OF OR OUT CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALOUTGS OUT THE
+ * SOFTWARE.
+ */
+#include "helpers.h"
+
+#if defined(BATCH_SIZE) && defined(DATA_TYPE)
+/** Calculate the space to batch conversion.
+ *
+ * @note Datatype should be given as a preprocessor argument using -DDATA_TYPE=type. e.g. -DDATA_TYPE=float
+ * @note The block shape tensor rank must be passed at compile time using -DBLOCK_SHAPE_DIM. e.g. -DBLOCK_SHAPE_DIM=2
+ *
+ * @param[in]  input_ptr                                 Pointer to the source tensor. Supported data types: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32
+ * @param[in]  input_stride_x                            Stride of the source tensor in X dimension (in bytes)
+ * @param[in]  input_step_x                              input_stride_x * number of elements along X processed per workitem(in bytes)
+ * @param[in]  input_stride_y                            Stride of the source image in Y dimension (in bytes)
+ * @param[in]  input_step_y                              input_stride_y * number of elements along Y processed per workitem(in bytes)
+ * @param[in]  input_stride_z                            Stride of the source tensor in Z dimension (in bytes)
+ * @param[in]  input_step_z                              input_stride_z * number of elements along Z processed per workitem(in bytes)
+ * @param[in]  input_offset_first_element_in_bytes       The offset of the first element in the first source image
+ * @param[in]  paddings_ptr                              Pointer to the second source image. Supported data types: S32
+ * @param[in]  paddings_stride_x                         Stride of the paddinds tensor in X dimension (in bytes)
+ * @param[in]  paddings_step_x                           paddings_stride_x * number of elements along X processed per workitem(in bytes)
+ * @param[in]  paddings_stride_y                         Stride of the paddinds tensor in Y dimension (in bytes)
+ * @param[in]  paddings_step_y                           paddings_stride_y * number of elements along Y processed per workitem(in bytes)
+ * @param[in]  paddingse_offset_first_element_in_bytes   The offset of the first element in the second source image
+ * @param[in]  block_shape_ptr                           Pointer to the block shape tensor. Supported data types: S32
+ * @param[in]  block_shape_stride_x                      Stride of the block shape tensor in X dimension (in bytes)
+ * @param[in]  block_shape_step_x                        block_shape_stride_x * number of elements along X processed per workitem(in bytes)
+ * @param[in]  block_shape_stride_y                      Stride of the block shape tensor in Y dimension (in bytes)
+ * @param[in]  block_shape_step_y                        block_shape_stride_y * number of elements along Y processed per workitem(in bytes)
+ * @param[in]  block_shape_offset_first_element_in_bytes The offset of the first element in the block shapetensor
+ * @param[in]  batch_id                                  The output tensor batch id
+ * @param[out] output_ptr                                Pointer to the destination tensor. Supported data types: same as @p input_ptr
+ * @param[in]  output_stride_x                           Stride of the destination tensor in X dimension (in bytes)
+ * @param[in]  output_step_x                             output_stride_x * number of elements along X processed per workitem(in bytes)
+ * @param[in]  output_stride_y                           Stride of the destination tensor in Y dimension (in bytes)
+ * @param[in]  output_step_y                             output_stride_y * number of elements along Y processed per workitem(in bytes)
+ * @param[in]  output_stride_z                           Stride of the destination tensor in Z dimension (in bytes)
+ * @param[in]  output_step_z                             output_stride_z * number of elements along Z processed per workitem(in bytes)
+ * @param[in]  output_offset_first_element_in_bytes      The offset of the first element in the destination image
+ */
+__kernel void space_to_batch(
+    TENSOR4D_DECLARATION(input),
+    IMAGE_DECLARATION(paddings),
+    VECTOR_DECLARATION(block_shape),
+    const int batch_id,
+    TENSOR3D_DECLARATION(output))
+{
+    Tensor4D in    = CONVERT_TO_TENSOR4D_STRUCT_NO_STEP(input, 0);
+    Image    pad   = CONVERT_TO_IMAGE_STRUCT_NO_STEP(paddings);
+    Vector   block = CONVERT_TO_VECTOR_STRUCT_NO_STEP(block_shape);
+    Tensor3D out   = CONVERT_TO_TENSOR3D_STRUCT(output);
+
+    const int PAD_LEFT_X  = *((__global int *)offset(&pad, 0, 0));
+    const int PAD_RIGHT_X = *((__global int *)offset(&pad, 1, 0));
+    const int PAD_LEFT_Y  = *((__global int *)offset(&pad, 0, 1));
+    const int PAD_RIGHT_Y = *((__global int *)offset(&pad, 1, 1));
+
+    int block_x = *((__global int *)vector_offset(&block, 0));
+    int block_y = *((__global int *)vector_offset(&block, 1));
+
+    const int out_x = get_global_id(0);
+    const int out_y = get_global_id(1);
+    const int z     = get_global_id(2);
+
+    if((out_x >= PAD_LEFT_X && out_x < WIDTH_OUT - PAD_RIGHT_X) && (out_y >= PAD_LEFT_Y && out_y < HEIGHT_OUT - PAD_RIGHT_Y))
+    {
+        const int r                      = (BATCH_SIZE / (block_x * block_y));
+        const int w                      = batch_id % r;
+        const int in_x                   = (out_x - PAD_LEFT_X) * block_x + (batch_id / r) % block_x;
+        const int in_y                   = (out_y - PAD_LEFT_Y) * block_y + (batch_id / r) / block_x;
+        *((__global DATA_TYPE *)out.ptr) = *((__global DATA_TYPE *)tensor4D_offset(&in, in_x, in_y, z, w));
+    }
+}
+#endif // defined(BATCH_SIZE) && defined(DATA_TYPE)
+
+#if defined(BATCH_SIZE) && defined(DATA_TYPE) && defined(BLOCK_SHAPE_X) && defined(BLOCK_SHAPE_Y) && defined(PAD_LEFT_X) && defined(PAD_RIGHT_X) && defined(PAD_LEFT_Y) && defined(PAD_RIGHT_Y)
+/** Calculate the space to batch conversion.
+ *
+ * @note Datatype should be given as a preprocessor argument using -DDATA_TYPE=type. e.g. -DDATA_TYPE=float
+ * @note The input tensor batch size must be passed at compile time using -DBATCH_SIZE. e.g. -DBATCH_SIZE=2
+ * @note The block shape x must be passed at compile time using -DBLOCK_SHAPE_X. e.g. -DBLOCK_SHAPE_X=2
+ * @note The block shape y must be passed at compile time using -DBLOCK_SHAPE_Y. e.g. -DBLOCK_SHAPE_Y=2
+ * @note The starting pad value of x must be passed at compile time using -DPAD_LEFT_X. e.g. -DPAD_LEFT_X=2
+ * @note The ending pad value of x must be passed at compile time using -DPAD_RIGHT_X. e.g. -DPAD_RIGHT_X=2
+ * @note The starting pad value of y must be passed at compile time using -DPAD_LEFT_Y. e.g. -DPAD_LEFT_Y=2
+ * @note The ending pad value of  y must be passed at compile time using -DPAD_RIGHT_Y. e.g. -DPAD_RIGHT_X=2
+ *
+ * @param[in]  input_ptr                            Pointer to the source tensor. Supported data types: U8/S8/QASYMM8/U16/S16/F16/U32/S32/F32
+ * @param[in]  input_stride_x                       Stride of the source tensor in X dimension (in bytes)
+ * @param[in]  input_step_x                         input_stride_x * number of elements along X processed per workitem(in bytes)
+ * @param[in]  input_stride_y                       Stride of the source image in Y dimension (in bytes)
+ * @param[in]  input_step_y                         input_stride_y * number of elements along Y processed per workitem(in bytes)
+ * @param[in]  input_stride_z                       Stride of the source tensor in Z dimension (in bytes)
+ * @param[in]  input_step_z                         input_stride_z * number of elements along Z processed per workitem(in bytes)
+ * @param[in]  input_offset_first_element_in_bytes  The offset of the first element in the first source image
+ * @param[in]  batch_id                             The output tensor batch id
+ * @param[out] output_ptr                           Pointer to the destination tensor. Supported data types: same as @p input_ptr
+ * @param[in]  output_stride_x                      Stride of the destination tensor in X dimension (in bytes)
+ * @param[in]  output_step_x                        output_stride_x * number of elements along X processed per workitem(in bytes)
+ * @param[in]  output_stride_y                      Stride of the destination tensor in Y dimension (in bytes)
+ * @param[in]  output_step_y                        output_stride_y * number of elements along Y processed per workitem(in bytes)
+ * @param[in]  output_stride_z                      Stride of the source tensor in Z dimension (in bytes)
+ * @param[in]  output_step_z                        output_stride_z * number of elements along Z processed per workitem(in bytes)
+ * @param[in]  output_offset_first_element_in_bytes The offset of the first element in the destination image
+ */
+__kernel void space_to_batch_static(
+    TENSOR4D_DECLARATION(input),
+    const int batch_id,
+    TENSOR3D_DECLARATION(output))
+{
+    Tensor4D in  = CONVERT_TO_TENSOR4D_STRUCT_NO_STEP(input, 0);
+    Tensor3D out = CONVERT_TO_TENSOR3D_STRUCT(output);
+
+    int block_x = BLOCK_SHAPE_X;
+    int block_y = *((__global int *)vector_offset(&block, 1));
+
+    const int out_x = get_global_id(0);
+    const int out_y = get_global_id(1);
+    const int z     = get_global_id(2);
+
+    if((out_x >= PAD_LEFT_X && out_x < WIDTH_OUT - PAD_RIGHT_X) && (out_y >= PAD_LEFT_Y && out_y < HEIGHT_OUT - PAD_RIGHT_Y))
+    {
+        const int r                      = (BATCH_SIZE / (block_x * block_y));
+        const int w                      = batch_id % r;
+        const int in_x                   = (out_x - PAD_LEFT_X) * block_x + (batch_id / r) % block_x;
+        const int in_y                   = (out_y - PAD_LEFT_Y) * block_y + (batch_id / r) / block_x;
+        *((__global DATA_TYPE *)out.ptr) = *((__global DATA_TYPE *)tensor4D_offset(&in, in_x, in_y, z, w));
+    }
+}
+#endif // defined(BATCH_SIZE) && defined(DATA_TYPE) && defined(BLOCK_SHAPE_X) && defined(BLOCK_SHAPE_Y) && defined(PAD_LEFT_X) && defined(PAD_RIGHT_X) && defined(PAD_LEFT_Y) && defined(PAD_RIGHT_Y)
diff --git a/src/core/CL/kernels/CLSpaceToBatchLayerKernel.cpp b/src/core/CL/kernels/CLSpaceToBatchLayerKernel.cpp
new file mode 100644
index 0000000..cda6e96
--- /dev/null
+++ b/src/core/CL/kernels/CLSpaceToBatchLayerKernel.cpp
@@ -0,0 +1,182 @@
+/*
+ * Copyright (c) 2018 ARM Limited.
+ *
+ * SPDX-License-Identifier: MIT
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a copy
+ * of this software and associated documentation files (the "Software"), to
+ * deal in the Software without restriction, including without limitation the
+ * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or
+ * sell copies of the Software, and to permit persons to whom the Software is
+ * furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in all
+ * copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+ * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+ * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+#include "arm_compute/core/CL/kernels/CLSpaceToBatchLayerKernel.h"
+
+#include "arm_compute/core/CL/CLHelpers.h"
+#include "arm_compute/core/CL/CLValidate.h"
+#include "arm_compute/core/CL/ICLTensor.h"
+#include "arm_compute/core/utils/misc/ShapeCalculator.h"
+
+using namespace arm_compute::misc::shape_calculator;
+
+namespace arm_compute
+{
+namespace
+{
+Status validate_arguments(const ITensorInfo *input, const ITensorInfo *block_info, const ITensorInfo *padddings, const ITensorInfo *output)
+{
+    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, block_info, padddings, output);
+    ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(block_info, 1, DataType::S32);
+    ARM_COMPUTE_RETURN_ERROR_ON(input->num_dimensions() > 4);
+
+    // Validate output if initialized
+    if(output->total_size() != 0)
+    {
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output);
+    }
+
+    return Status{};
+}
+Status validate_arguments_static(const ITensorInfo *input, const int block_shape_x, const int block_shape_y, const Size2D &padding_left, const Size2D &padding_right,
+                                 const ITensorInfo *output)
+{
+    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, output);
+    ARM_COMPUTE_RETURN_ERROR_ON(block_shape_x < 1 || block_shape_y < 1);
+    ARM_COMPUTE_RETURN_ERROR_ON(input->num_dimensions() > 4);
+
+    // Validate output if initialized
+    if(output->total_size() != 0)
+    {
+        ARM_COMPUTE_RETURN_ERROR_ON(output->tensor_shape()[0] < padding_left.x() + padding_right.y());
+        ARM_COMPUTE_RETURN_ERROR_ON(input->tensor_shape()[0] / block_shape_x != (output->tensor_shape()[0] - padding_left.x() - padding_right.y()));
+        ARM_COMPUTE_RETURN_ERROR_ON(input->tensor_shape()[1] / block_shape_y != (output->tensor_shape()[1] - padding_left.x() - padding_right.y()));
+        ARM_COMPUTE_RETURN_ERROR_ON(input->tensor_shape()[2] != output->tensor_shape()[2]);
+        ARM_COMPUTE_RETURN_ERROR_ON(output->tensor_shape()[3] % (block_shape_x * block_shape_y) != 0);
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output);
+    }
+
+    return Status{};
+}
+} // namespace
+
+CLSpaceToBatchLayerKernel::CLSpaceToBatchLayerKernel()
+    : _input(nullptr), _block_shape(nullptr), _paddings(nullptr), _output(nullptr)
+{
+}
+
+void CLSpaceToBatchLayerKernel::configure(const ICLTensor *input, const ICLTensor *block_shape, const ICLTensor *paddings, ICLTensor *output)
+{
+    ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
+    ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), block_shape->info(), paddings->info(), output->info()));
+
+    _input       = input;
+    _block_shape = block_shape;
+    _paddings    = paddings;
+    _output      = output;
+
+    // Create kernel
+    CLBuildOptions build_opts;
+    build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(input->info()->data_type()));
+    build_opts.add_option("-DWIDTH_OUT=" + support::cpp11::to_string(output->info()->dimension(0)));
+    build_opts.add_option("-DHEIGHT_OUT=" + support::cpp11::to_string(output->info()->dimension(1)));
+    build_opts.add_option("-DBATCH_SIZE=" + support::cpp11::to_string(output->info()->dimension(3)));
+    _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("space_to_batch", build_opts.options()));
+
+    // Configure kernel window
+    Window win = calculate_max_window(*output->info(), Steps());
+    ICLKernel::configure_internal(win);
+}
+
+void CLSpaceToBatchLayerKernel::configure(const ICLTensor *input, const int block_shape_x, const int block_shape_y, const Size2D &padding_left, const Size2D &padding_right,
+                                          ICLTensor *output)
+{
+    ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
+
+    TensorShape output_shape = misc::shape_calculator::compute_space_to_batch_shape(input->info(), block_shape_x, block_shape_y, padding_left, padding_right);
+    auto_init_if_empty(*output->info(), output_shape, 1, input->info()->data_type());
+
+    ARM_COMPUTE_ERROR_THROW_ON(validate_arguments_static(input->info(), block_shape_x, block_shape_y, padding_left, padding_right, output->info()));
+
+    _input  = input;
+    _output = output;
+
+    // Create kernel
+    CLBuildOptions build_opts;
+    build_opts.add_option("-DDATA_TYPE=" + get_cl_type_from_data_type(input->info()->data_type()));
+    build_opts.add_option("-DWIDTH_OUT=" + support::cpp11::to_string(output->info()->dimension(0)));
+    build_opts.add_option("-DHEIGHT_OUT=" + support::cpp11::to_string(output->info()->dimension(1)));
+    build_opts.add_option("-DBATCH_SIZE=" + support::cpp11::to_string(output->info()->dimension(3)));
+    build_opts.add_option("-DBLOCK_SHAPE_X=" + support::cpp11::to_string(block_shape_x));
+    build_opts.add_option("-DBLOCK_SHAPE_Y=" + support::cpp11::to_string(block_shape_y));
+    build_opts.add_option("-DPAD_START_X=" + support::cpp11::to_string(padding_left.x()));
+    build_opts.add_option("-DPAD_END_X=" + support::cpp11::to_string(padding_right.x()));
+    build_opts.add_option("-DPAD_START_Y=" + support::cpp11::to_string(padding_left.y()));
+    build_opts.add_option("-DPAD_END_Y=" + support::cpp11::to_string(padding_right.y()));
+    _kernel = static_cast<cl::Kernel>(CLKernelLibrary::get().create_kernel("space_to_batch_static", build_opts.options()));
+
+    // Configure kernel window
+    Window win = calculate_max_window(*output->info(), Steps());
+    ICLKernel::configure_internal(win);
+}
+
+Status CLSpaceToBatchLayerKernel::validate(const ITensorInfo *input, const ITensorInfo *block_shape, const ITensorInfo *paddings, const ITensorInfo *output)
+{
+    ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, block_shape, paddings, output));
+    return Status{};
+}
+Status CLSpaceToBatchLayerKernel::validate(const ITensorInfo *input, const int block_shape_x, const int block_shape_y, const Size2D &padding_left, const Size2D &padding_right,
+                                           const ITensorInfo *output)
+{
+    ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments_static(input, block_shape_x, block_shape_y, padding_left, padding_right, output));
+    return Status{};
+}
+
+void CLSpaceToBatchLayerKernel::run(const Window &window, cl::CommandQueue &queue)
+{
+    ARM_COMPUTE_ERROR_ON_UNCONFIGURED_KERNEL(this);
+    ARM_COMPUTE_ERROR_ON_INVALID_SUBWINDOW(ICLKernel::window(), window);
+
+    Window slice_out = window.first_slice_window_3D();
+
+    Window slice_in = window.first_slice_window_4D();
+    slice_in.set(Window::DimX, Window::Dimension(0, 0, 0));
+    slice_in.set(Window::DimY, Window::Dimension(0, 0, 0));
+    slice_in.set(Window::DimZ, Window::Dimension(0, 0, 0));
+    slice_in.set(3, Window::Dimension(0, 0, 0));
+
+    Window vector_slice = window.first_slice_window_1D();
+    vector_slice.set(Window::DimX, Window::Dimension(0, 0, 0));
+
+    Window padding_slice = window.first_slice_window_2D();
+    padding_slice.set(Window::DimX, Window::Dimension(0, 0, 0));
+    padding_slice.set(Window::DimY, Window::Dimension(0, 0, 0));
+
+    int batch_id = 0;
+    do
+    {
+        unsigned int idx = 0;
+        add_4D_tensor_argument(idx, _input, slice_in);
+        if(_paddings != nullptr && _block_shape != nullptr)
+        {
+            add_2D_tensor_argument(idx, _paddings, padding_slice);
+            add_1D_tensor_argument(idx, _block_shape, vector_slice);
+        }
+        add_argument(idx, batch_id);
+        add_3D_tensor_argument(idx, _output, slice_out);
+        enqueue(queue, *this, slice_out);
+        ++batch_id;
+    }
+    while(window.slide_window_slice_3D(slice_out));
+}
+} // namespace arm_compute
\ No newline at end of file