diff --git a/src/core/CL/CLKernelLibrary.cpp b/src/core/CL/CLKernelLibrary.cpp
index 9dec79b..0b59ec8 100644
--- a/src/core/CL/CLKernelLibrary.cpp
+++ b/src/core/CL/CLKernelLibrary.cpp
@@ -283,6 +283,7 @@
     { "lktracker_stage0", "optical_flow_pyramid_lk.cl" },
     { "lktracker_stage1", "optical_flow_pyramid_lk.cl" },
     { "magnitude_phase", "magnitude_phase.cl" },
+    { "max_unpooling_layer_2", "unpooling_layer.cl" },
     { "mean_stddev_accumulate", "mean_stddev.cl" },
     { "mean_stddev_normalization", "mean_stddev_normalization.cl" },
     { "memset", "memset.cl" },
@@ -939,6 +940,10 @@
 #include "./cl_kernels/types.hembed"
     },
     {
+        "unpooling_layer.cl",
+#include "./cl_kernels/unpooling_layer.clembed"
+    },
+    {
         "warp_affine.cl",
 #include "./cl_kernels/warp_affine.clembed"
     },
diff --git a/src/core/CL/cl_kernels/helpers.h b/src/core/CL/cl_kernels/helpers.h
index d3ede32..7b08233 100644
--- a/src/core/CL/cl_kernels/helpers.h
+++ b/src/core/CL/cl_kernels/helpers.h
@@ -516,6 +516,10 @@
 #define CONVERT_TO_TENSOR4D_STRUCT_NO_STEP(name, mod_size) \
     update_tensor4D_workitem_ptr(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, 0, name##_stride_y, 0, name##_stride_z, 0, name##_stride_w, 0, mod_size)
 
+#define CONVERT_TO_TENSOR3D_STRUCT_NO_UPDATE_PTR(name)                                                                                       \
+    tensor3D_ptr_no_update(name##_ptr, name##_offset_first_element_in_bytes, name##_stride_x, name##_step_x, name##_stride_y, name##_step_y, \
+                           name##_stride_z, name##_step_z)
+
 /** Structure to hold Vector information */
 typedef struct Vector
 {
@@ -652,6 +656,32 @@
     return tensor;
 }
 
+/** Wrap 3D tensor information into an tensor structure.
+ *
+ * @param[in] ptr                           Pointer to the starting postion of the buffer
+ * @param[in] offset_first_element_in_bytes The offset of the first element in the source image
+ * @param[in] stride_x                      Stride of the image in X dimension (in bytes)
+ * @param[in] step_x                        stride_x * number of elements along X processed per workitem(in bytes)
+ * @param[in] stride_y                      Stride of the image in Y dimension (in bytes)
+ * @param[in] step_y                        stride_y * number of elements along Y processed per workitem(in bytes)
+ * @param[in] stride_z                      Stride of the image in Z dimension (in bytes)
+ * @param[in] step_z                        stride_z * number of elements along Z processed per workitem(in bytes)
+ *
+ * @return A 3D tensor object
+ */
+inline Tensor3D tensor3D_ptr_no_update(__global uchar *ptr, uint offset_first_element_in_bytes, uint stride_x, uint step_x, uint stride_y, uint step_y, uint stride_z, uint step_z)
+{
+    Tensor3D tensor =
+    {
+        .ptr                           = ptr,
+        .offset_first_element_in_bytes = offset_first_element_in_bytes,
+        .stride_x                      = stride_x,
+        .stride_y                      = stride_y,
+        .stride_z                      = stride_z
+    };
+    return tensor;
+}
+
 inline Tensor4D update_tensor4D_workitem_ptr(__global uchar *ptr, uint offset_first_element_in_bytes, uint stride_x, uint step_x, uint stride_y, uint step_y, uint stride_z, uint step_z, uint stride_w,
                                              uint step_w,
                                              uint mod_size)
@@ -716,4 +746,29 @@
     return tensor->ptr + x * tensor->stride_x + y * tensor->stride_y + z * tensor->stride_z + w * tensor->stride_w;
 }
 
+/** Get the offset for a given linear index of a Tensor3D
+ *
+ * @param[in] tensor Pointer to the starting position of the buffer
+ * @param[in] width  Width of the input tensor
+ * @param[in] height Height of the input tensor
+ * @param[in] depth  Depth of the input tensor
+ * @param[in] index  Linear index
+ */
+inline __global const uchar *tensor3D_index2ptr(const Tensor3D *tensor, uint width, uint height, uint depth, uint index)
+{
+    uint num_elements = width * height;
+
+    const uint z = index / num_elements;
+
+    index %= num_elements;
+
+    const uint y = index / width;
+
+    index %= width;
+
+    const uint x = index;
+
+    return tensor->ptr + x * tensor->stride_x + y * tensor->stride_y + z * tensor->stride_z + tensor->offset_first_element_in_bytes;
+}
+
 #endif // _HELPER_H
diff --git a/src/core/CL/cl_kernels/unpooling_layer.cl b/src/core/CL/cl_kernels/unpooling_layer.cl
new file mode 100644
index 0000000..457e9bf
--- /dev/null
+++ b/src/core/CL/cl_kernels/unpooling_layer.cl
@@ -0,0 +1,72 @@
+/*
+ * Copyright (c) 2020 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 "helpers.h"
+
+/** Performs max unpooling function with pool size equal to 2.
+ *
+ * @note Datatype must be passed using -DDATA_TYPE e.g. -DDATA_TYPE=float. Supported data types are F16/F32
+ * @note The width of the output tensor must be passed using -DWIDTH_DST e.g. -DWIDTH_DST=24
+ * @note The height of the output tensor must be passed using -DHEIGHT_DST e.g. -DHEIGHT_DST=54
+ * @note The depth of the output tensor must be passed using -DDEPTH_DST e.g. -DDEPTH_DST=32
+ *
+ * @param[in]  input_ptr                             Pointer to the source tensor. Supported data types: F16/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 tensor 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 source tensor
+ * @param[out] output_ptr                            Pointer to the output tensor. Supported data types: same as @p input_ptr
+ * @param[in]  output_stride_x                       Stride of the output 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 output 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 output 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 output tensor
+ * @param[in]  indices_ptr                           Pointer to the indices tensor. Supported data types: U32
+ * @param[in]  indices_stride_x                      Stride of the indices tensor in X dimension (in bytes)
+ * @param[in]  indices_step_x                        indices_stride_x * number of elements along X processed per workitem(in bytes)
+ * @param[in]  indices_stride_y                      Stride of the indices tensor in Y dimension (in bytes)
+ * @param[in]  indices_step_y                        indices_stride_y * number of elements along Y processed per workitem(in bytes)
+ * @param[in]  indices_stride_z                      Stride of the indices tensor in Z dimension (in bytes)
+ * @param[in]  indices_step_z                        indices_stride_z * number of elements along Z processed per workitem(in bytes)
+ * @param[in]  indices_offset_first_element_in_bytes The offset of the first element in the indices tensor
+ */
+__kernel void max_unpooling_layer_2(
+    TENSOR3D_DECLARATION(input),
+    TENSOR3D_DECLARATION(output),
+    TENSOR3D_DECLARATION(indices))
+{
+    Tensor3D input   = CONVERT_TO_TENSOR3D_STRUCT(input);
+    Tensor3D output  = CONVERT_TO_TENSOR3D_STRUCT_NO_UPDATE_PTR(output);
+    Tensor3D indices = CONVERT_TO_TENSOR3D_STRUCT(indices);
+
+    unsigned int index = *((__global unsigned int *)indices.ptr);
+    DATA_TYPE value    = *((__global DATA_TYPE *)input.ptr);
+
+    *((__global DATA_TYPE *)tensor3D_index2ptr(&output, WIDTH_DST, HEIGHT_DST, DEPTH_DST, index)) = value;
+}
\ No newline at end of file
diff --git a/src/core/CL/kernels/CLMaxUnpoolingLayerKernel.cpp b/src/core/CL/kernels/CLMaxUnpoolingLayerKernel.cpp
new file mode 100644
index 0000000..08c7464
--- /dev/null
+++ b/src/core/CL/kernels/CLMaxUnpoolingLayerKernel.cpp
@@ -0,0 +1,145 @@
+/*
+ * Copyright (c) 2020 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/CLMaxUnpoolingLayerKernel.h"
+
+#include "arm_compute/core/AccessWindowStatic.h"
+#include "arm_compute/core/CL/CLHelpers.h"
+#include "arm_compute/core/CL/CLKernelLibrary.h"
+#include "arm_compute/core/CL/CLValidate.h"
+#include "arm_compute/core/CL/ICLTensor.h"
+#include "arm_compute/core/CL/OpenCL.h"
+#include "arm_compute/core/Error.h"
+#include "arm_compute/core/Helpers.h"
+#include "arm_compute/core/TensorInfo.h"
+#include "arm_compute/core/Types.h"
+#include "arm_compute/core/Utils.h"
+#include "arm_compute/core/Window.h"
+#include "arm_compute/core/utils/misc/ShapeCalculator.h"
+#include "support/StringSupport.h"
+
+namespace arm_compute
+{
+using namespace misc::shape_calculator;
+
+namespace
+{
+Status validate_arguments(const ITensorInfo *input, const ITensorInfo *output, const PoolingLayerInfo &pool_info, const ITensorInfo *indices)
+{
+    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, output, indices);
+    ARM_COMPUTE_RETURN_ERROR_ON_F16_UNSUPPORTED(input);
+    ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(input, 1, DataType::QASYMM8, DataType::QASYMM8_SIGNED, DataType::F16, DataType::F32);
+    ARM_COMPUTE_RETURN_ERROR_ON_DATA_TYPE_CHANNEL_NOT_IN(indices, 1, DataType::U32);
+    ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_SHAPES(input, indices);
+
+    int                 pool_stride_x   = 0;
+    int                 pool_stride_y   = 0;
+    PoolingType         pool_type       = pool_info.pool_type;
+    const PadStrideInfo pad_stride_info = pool_info.pad_stride_info;
+    std::tie(pool_stride_x, pool_stride_y) = pad_stride_info.stride();
+    const int    pool_size_x = pool_info.pool_size.width;
+    const int    pool_size_y = pool_info.pool_size.height;
+    const Size2D pool_size(pool_size_x, pool_size_y);
+
+    ARM_COMPUTE_RETURN_ERROR_ON_MSG(pool_type != PoolingType::MAX, "Pooling indices only supported for MAX pooling method");
+    ARM_COMPUTE_RETURN_ERROR_ON_MSG((pool_size != Size2D(2, 2)), "Pooling indices only supported for pool size 2x2");
+    if(output->total_size() != 0)
+    {
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_TYPES(input, output);
+        ARM_COMPUTE_RETURN_ERROR_ON_MISMATCHING_DATA_LAYOUT(input, output);
+    }
+
+    return Status{};
+}
+} // namespace
+
+CLMaxUnpoolingLayerKernel::CLMaxUnpoolingLayerKernel()
+    : _input(nullptr), _output(nullptr), _indices(nullptr)
+{
+}
+
+void CLMaxUnpoolingLayerKernel::configure(const CLCompileContext &compile_context, const ICLTensor *input, const ICLTensor *indices, ICLTensor *output, const PoolingLayerInfo &pool_info)
+{
+    ARM_COMPUTE_ERROR_ON_NULLPTR(input, output);
+    ARM_COMPUTE_ERROR_THROW_ON(validate_arguments(input->info(), output->info(), pool_info, indices->info()));
+
+    _input   = input;
+    _output  = output;
+    _indices = indices;
+
+    // Create build options
+    CLBuildOptions build_opts;
+    build_opts.add_option("-DDATA_TYPE=" + get_cl_unsigned_type_from_element_size(input->info()->element_size()));
+    build_opts.add_option("-DWIDTH_DST=" + support::cpp11::to_string(output->info()->dimension(0)));
+    build_opts.add_option("-DHEIGHT_DST=" + support::cpp11::to_string(output->info()->dimension(1)));
+    build_opts.add_option("-DDEPTH_DST=" + support::cpp11::to_string(output->info()->dimension(2)));
+
+    const std::string kernel_name("max_unpooling_layer_2");
+
+    // Create kernel
+    _kernel = create_kernel(compile_context, kernel_name, build_opts.options());
+
+    const TensorShape output_shape = compute_unpool_shape(*input->info(), pool_info);
+    auto_init_if_empty(*output->info(), input->info()->clone()->set_tensor_shape(output_shape));
+
+    auto window = calculate_max_window(*input->info(), Steps());
+    ICLKernel::configure_internal(window);
+
+    // Set config_id for enabling LWS tuning
+    _config_id = kernel_name;
+    _config_id += lower_string(string_from_data_type(input->info()->data_type()));
+    _config_id += "_";
+    _config_id += support::cpp11::to_string(output->info()->dimension(0));
+    _config_id += "_";
+    _config_id += support::cpp11::to_string(output->info()->dimension(1));
+    _config_id += "_";
+    _config_id += support::cpp11::to_string(output->info()->dimension(2));
+    _config_id += "_";
+    _config_id += support::cpp11::to_string(output->info()->dimension(3));
+}
+
+Status CLMaxUnpoolingLayerKernel::validate(const ITensorInfo *input, const ITensorInfo *indices, const ITensorInfo *output, const PoolingLayerInfo &pool_info)
+{
+    ARM_COMPUTE_RETURN_ERROR_ON_NULLPTR(input, indices, output);
+    ARM_COMPUTE_RETURN_ON_ERROR(validate_arguments(input, output, pool_info, indices));
+    return Status{};
+}
+
+void CLMaxUnpoolingLayerKernel::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 = window.first_slice_window_3D();
+
+    do
+    {
+        unsigned int idx = 0;
+        add_3D_tensor_argument(idx, _input, slice);
+        add_3D_tensor_argument(idx, _output, slice);
+        add_3D_tensor_argument(idx, _indices, slice);
+        enqueue(queue, *this, slice, lws_hint());
+    }
+    while(window.slide_window_slice_3D(slice));
+}
+} // namespace arm_compute
diff --git a/src/runtime/CL/functions/CLMaxUnpoolingLayer.cpp b/src/runtime/CL/functions/CLMaxUnpoolingLayer.cpp
new file mode 100644
index 0000000..3e32c55
--- /dev/null
+++ b/src/runtime/CL/functions/CLMaxUnpoolingLayer.cpp
@@ -0,0 +1,65 @@
+/*
+ * Copyright (c) 2020 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/runtime/CL/functions/CLMaxUnpoolingLayer.h"
+
+#include "arm_compute/core/CL/ICLTensor.h"
+#include "arm_compute/core/CL/kernels/CLMaxUnpoolingLayerKernel.h"
+#include "arm_compute/core/PixelValue.h"
+#include "arm_compute/core/Validate.h"
+#include "arm_compute/runtime/CL/CLScheduler.h"
+
+namespace arm_compute
+{
+CLMaxUnpoolingLayer::CLMaxUnpoolingLayer()
+    : _memset_kernel(), _unpooling_layer_kernel()
+{
+}
+
+void CLMaxUnpoolingLayer::configure(ICLTensor *input, ICLTensor *indices, ICLTensor *output, const PoolingLayerInfo &pool_info)
+{
+    configure(CLKernelLibrary::get().get_compile_context(), input, indices, output, pool_info);
+}
+
+void CLMaxUnpoolingLayer::configure(const CLCompileContext &compile_context, ICLTensor *input, ICLTensor *indices, ICLTensor *output, const PoolingLayerInfo &pool_info)
+{
+    const PixelValue zero_value(0.f);
+    _memset_kernel.configure(output, zero_value);
+
+    _unpooling_layer_kernel.configure(compile_context, input, indices, output, pool_info);
+}
+
+Status CLMaxUnpoolingLayer::validate(const ITensorInfo *input, const ITensorInfo *indices, const ITensorInfo *output, const PoolingLayerInfo &pool_info)
+{
+    return CLMaxUnpoolingLayerKernel::validate(input, indices, output, pool_info);
+}
+
+void CLMaxUnpoolingLayer::run()
+{
+    // Run memset
+    CLScheduler::get().enqueue(_memset_kernel, false);
+
+    // Run max unpooling layer
+    CLScheduler::get().enqueue(_unpooling_layer_kernel);
+}
+} /* namespace arm_compute */