COMPMID-: Implement L2NormalizationLayer.

Change-Id: I2223850b84826a87650729d287c57c22bcab7462
Reviewed-on: http://mpd-gerrit.cambridge.arm.com/82767
Tested-by: Kaizen <jeremy.johnson+kaizengerrit@arm.com>
Reviewed-by: Anthony Barbier <anthony.barbier@arm.com>
diff --git a/arm_compute/core/NEON/NEKernels.h b/arm_compute/core/NEON/NEKernels.h
index a326f36..e49d04a 100644
--- a/arm_compute/core/NEON/NEKernels.h
+++ b/arm_compute/core/NEON/NEKernels.h
@@ -70,6 +70,7 @@
 #include "arm_compute/core/NEON/kernels/NEHistogramKernel.h"
 #include "arm_compute/core/NEON/kernels/NEIm2ColKernel.h"
 #include "arm_compute/core/NEON/kernels/NEIntegralImageKernel.h"
+#include "arm_compute/core/NEON/kernels/NEL2NormalizeKernel.h"
 #include "arm_compute/core/NEON/kernels/NELKTrackerKernel.h"
 #include "arm_compute/core/NEON/kernels/NELocallyConnectedMatrixMultiplyKernel.h"
 #include "arm_compute/core/NEON/kernels/NEMagnitudePhaseKernel.h"
@@ -82,6 +83,7 @@
 #include "arm_compute/core/NEON/kernels/NEPixelWiseMultiplicationKernel.h"
 #include "arm_compute/core/NEON/kernels/NEPoolingLayerKernel.h"
 #include "arm_compute/core/NEON/kernels/NEROIPoolingLayerKernel.h"
+#include "arm_compute/core/NEON/kernels/NEReductionOperationKernel.h"
 #include "arm_compute/core/NEON/kernels/NERemapKernel.h"
 #include "arm_compute/core/NEON/kernels/NEScaleKernel.h"
 #include "arm_compute/core/NEON/kernels/NEScharr3x3Kernel.h"
diff --git a/arm_compute/core/NEON/kernels/NEL2NormalizeKernel.h b/arm_compute/core/NEON/kernels/NEL2NormalizeKernel.h
new file mode 100644
index 0000000..545aeaa
--- /dev/null
+++ b/arm_compute/core/NEON/kernels/NEL2NormalizeKernel.h
@@ -0,0 +1,70 @@
+/*
+ * Copyright (c) 2017 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.
+ */
+#ifndef __ARM_COMPUTE_NEL2NORMALIZEKERNEL_H__
+#define __ARM_COMPUTE_NEL2NORMALIZEKERNEL_H__
+
+#include "arm_compute/core/NEON/INEKernel.h"
+
+namespace arm_compute
+{
+class ITensor;
+
+/** Interface for performing a L2 normalize on a given axis given the square sum of it in this axis */
+class NEL2NormalizeKernel : public INEKernel
+{
+public:
+    /** Default constructor */
+    NEL2NormalizeKernel();
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    NEL2NormalizeKernel(const NEL2NormalizeKernel &) = delete;
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    NEL2NormalizeKernel &operator=(const NEL2NormalizeKernel &) = delete;
+    /** Allow instances of this class to be moved */
+    NEL2NormalizeKernel(NEL2NormalizeKernel &&) = default;
+    /** Allow instances of this class to be moved */
+    NEL2NormalizeKernel &operator=(NEL2NormalizeKernel &&) = default;
+    /** Default destructor */
+    ~NEL2NormalizeKernel() = default;
+    /** Set the input and output tensors.
+     *
+     * @param[in]  input   Source tensor. Data types supported: F32.
+     * @param[in]  sum     Sum values tensor. Data types supported: same as @p input.
+     * @param[out] output  Destination tensor. Data types supported: same as @p input.
+     * @param[in]  axis    Dimension along which to reduce. Supported reduction axis : 0
+     * @param[in]  epsilon Lower bound value for the normalization.
+     */
+    void configure(const ITensor *input, const ITensor *sum, ITensor *output, unsigned int axis, float epsilon);
+
+    // Inherited methods overridden:
+    void run(const Window &window) override;
+
+private:
+    const ITensor *_input;
+    const ITensor *_sum;
+    ITensor       *_output;
+    unsigned int   _axis;
+    float          _epsilon;
+};
+}
+#endif /*__ARM_COMPUTE_NEL2NORMALIZEKERNEL_H__ */
diff --git a/arm_compute/core/NEON/kernels/NEReductionOperationKernel.h b/arm_compute/core/NEON/kernels/NEReductionOperationKernel.h
new file mode 100644
index 0000000..147e835
--- /dev/null
+++ b/arm_compute/core/NEON/kernels/NEReductionOperationKernel.h
@@ -0,0 +1,69 @@
+/*
+ * Copyright (c) 2017 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.
+ */
+#ifndef __ARM_COMPUTE_NEREDUCTIONOPERATIONKERNEL_H__
+#define __ARM_COMPUTE_NEREDUCTIONOPERATIONKERNEL_H__
+
+#include "arm_compute/core/NEON/INEKernel.h"
+
+namespace arm_compute
+{
+class ITensor;
+
+/** NEON kernel to perform a reduction operation */
+class NEReductionOperationKernel : public INEKernel
+{
+public:
+    /** Default constructor */
+    NEReductionOperationKernel();
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    NEReductionOperationKernel(const NEReductionOperationKernel &) = delete;
+    /** Prevent instances of this class from being copied (As this class contains pointers) */
+    NEReductionOperationKernel &operator=(const NEReductionOperationKernel &) = delete;
+    /** Allow instances of this class to be moved */
+    NEReductionOperationKernel(NEReductionOperationKernel &&) = default;
+    /** Allow instances of this class to be moved */
+    NEReductionOperationKernel &operator=(NEReductionOperationKernel &&) = default;
+    /** Default destructor */
+    ~NEReductionOperationKernel() = default;
+    /** Set the source, destination of the kernel
+     *
+     * @param[in]  input  Source tensor. Data type supported: F32.
+     * @param[out] output Destination tensor.Data types supported: same as @p input.
+     * @param[in]  axis   Axis along which to reduce. Supported reduction axis : 0
+     * @param[in]  op     Reduction operation to perform.
+     */
+    void configure(const ITensor *input, ITensor *output, unsigned int axis, ReductionOperation op);
+    // Inherited methods overridden:
+    void run(const Window &window) override;
+    BorderSize border_size() const override;
+
+private:
+    const ITensor     *_input;
+    ITensor           *_output;
+    unsigned int       _reduction_axis;
+    ReductionOperation _op;
+    BorderSize         _border_size;
+};
+}
+#endif /*__ARM_COMPUTE_NEREDUCTIONOPERATIONKERNEL_H__ */
diff --git a/arm_compute/core/Types.h b/arm_compute/core/Types.h
index 2d3b3d6..3092976 100644
--- a/arm_compute/core/Types.h
+++ b/arm_compute/core/Types.h
@@ -342,6 +342,12 @@
     MAX    = 2, /**< Non linear dilate. */
 };
 
+/** Available reduction operations */
+enum class ReductionOperation
+{
+    SUM_SQUARE, /**< Sum of squares */
+};
+
 /** The normalization type used for the normalization layer */
 enum class NormType
 {