Anthony Barbier | 6ff3b19 | 2017-09-04 18:44:23 +0100 | [diff] [blame] | 1 | /* |
| 2 | * Copyright (c) 2016, 2017 ARM Limited. |
| 3 | * |
| 4 | * SPDX-License-Identifier: MIT |
| 5 | * |
| 6 | * Permission is hereby granted, free of charge, to any person obtaining a copy |
| 7 | * of this software and associated documentation files (the "Software"), to |
| 8 | * deal in the Software without restriction, including without limitation the |
| 9 | * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or |
| 10 | * sell copies of the Software, and to permit persons to whom the Software is |
| 11 | * furnished to do so, subject to the following conditions: |
| 12 | * |
| 13 | * The above copyright notice and this permission notice shall be included in all |
| 14 | * copies or substantial portions of the Software. |
| 15 | * |
| 16 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
| 17 | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| 18 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE |
| 19 | * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER |
| 20 | * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, |
| 21 | * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE |
| 22 | * SOFTWARE. |
| 23 | */ |
| 24 | #ifndef __ARM_COMPUTE_NECHANNELCOMBINEKERNEL_H__ |
| 25 | #define __ARM_COMPUTE_NECHANNELCOMBINEKERNEL_H__ |
| 26 | |
| 27 | #include "arm_compute/core/NEON/INEKernel.h" |
| 28 | |
| 29 | #include <array> |
| 30 | #include <cstdint> |
| 31 | |
| 32 | namespace arm_compute |
| 33 | { |
| 34 | class IMultiImage; |
| 35 | class ITensor; |
| 36 | using IImage = ITensor; |
| 37 | |
| 38 | /** Interface for the channel combine kernel */ |
| 39 | class NEChannelCombineKernel : public INEKernel |
| 40 | { |
| 41 | public: |
| 42 | /** Default constructor */ |
| 43 | NEChannelCombineKernel(); |
| 44 | /** Prevent instances of this class from being copied (As this class contains pointers) */ |
| 45 | NEChannelCombineKernel(const NEChannelCombineKernel &) = delete; |
| 46 | /** Prevent instances of this class from being copied (As this class contains pointers) */ |
| 47 | NEChannelCombineKernel &operator=(const NEChannelCombineKernel &) = delete; |
| 48 | /** Allow instances of this class to be moved */ |
| 49 | NEChannelCombineKernel(NEChannelCombineKernel &&) = default; |
| 50 | /** Allow instances of this class to be moved */ |
| 51 | NEChannelCombineKernel &operator=(NEChannelCombineKernel &&) = default; |
| 52 | /** Default destructor */ |
| 53 | ~NEChannelCombineKernel() = default; |
| 54 | |
| 55 | /** Configure function's inputs and outputs. |
| 56 | * |
| 57 | * @param[in] plane0 The 2D plane that forms channel 0. Data type supported: U8 |
| 58 | * @param[in] plane1 The 2D plane that forms channel 1. Data type supported: U8 |
| 59 | * @param[in] plane2 The 2D plane that forms channel 2. Data type supported: U8 |
| 60 | * @param[in] plane3 The 2D plane that forms channel 3. Data type supported: U8 |
| 61 | * @param[out] output The single planar output tensor. Formats supported: RGB888/RGBA8888/UYVY422/YUYV422 |
| 62 | */ |
| 63 | void configure(const ITensor *plane0, const ITensor *plane1, const ITensor *plane2, const ITensor *plane3, ITensor *output); |
| 64 | /** Configure function's inputs and outputs. |
| 65 | * |
| 66 | * @param[in] plane0 The 2D plane that forms channel 0. Data type supported: U8 |
| 67 | * @param[in] plane1 The 2D plane that forms channel 1. Data type supported: U8 |
| 68 | * @param[in] plane2 The 2D plane that forms channel 2. Data type supported: U8 |
| 69 | * @param[out] output The multi planar output tensor. Formats supported: NV12/NV21/IYUV/YUV444 |
| 70 | */ |
| 71 | void configure(const IImage *plane0, const IImage *plane1, const IImage *plane2, IMultiImage *output); |
| 72 | |
| 73 | // Inherited methods overridden: |
Moritz Pflanzer | c186b57 | 2017-09-07 09:48:04 +0100 | [diff] [blame] | 74 | void run(const Window &window, const ThreadInfo &info) override; |
Anthony Barbier | 6ff3b19 | 2017-09-04 18:44:23 +0100 | [diff] [blame] | 75 | bool is_parallelisable() const override; |
| 76 | |
| 77 | private: |
| 78 | /** Combine 3 planes to form a three channel single plane tensor. |
| 79 | * |
| 80 | * @param[in] win Region on which to execute the kernel. |
| 81 | */ |
| 82 | void combine_3C(const Window &win); |
| 83 | /** Combine 4 planes to form a four channel single plane tensor. |
| 84 | * |
| 85 | * @param[in] win Region on which to execute the kernel. |
| 86 | */ |
| 87 | void combine_4C(const Window &win); |
| 88 | /** Combine 3 planes to form a single plane YUV tensor. |
| 89 | * |
| 90 | * @param[in] win Region on which to execute the kernel. |
| 91 | */ |
| 92 | template <bool is_yuyv> |
| 93 | void combine_YUV_1p(const Window &win); |
| 94 | /** Combine 3 planes to form a two plane YUV tensor. |
| 95 | * |
| 96 | * @param[in] win Region on which to execute the kernel. |
| 97 | */ |
| 98 | void combine_YUV_2p(const Window &win); |
| 99 | /** Combine 3 planes to form a three plane YUV tensor. |
| 100 | * |
| 101 | * @param[in] win Region on which to execute the kernel. |
| 102 | */ |
| 103 | void combine_YUV_3p(const Window &win); |
| 104 | /** Copies a full plane to the output tensor. |
| 105 | * |
| 106 | * @param[in] win Region on which to execute the kernel. |
| 107 | */ |
| 108 | void copy_plane(const Window &win, uint32_t plane_id); |
| 109 | /** Common signature for all the specialised ChannelCombine functions |
| 110 | * |
| 111 | * @param[in] window Region on which to execute the kernel. |
| 112 | */ |
| 113 | using ChannelCombineFunction = void (NEChannelCombineKernel::*)(const Window &window); |
| 114 | /** ChannelCombine function to use for the particular tensor types passed to configure() */ |
| 115 | ChannelCombineFunction _func; |
| 116 | std::array<const ITensor *, 4> _planes; |
| 117 | ITensor *_output; |
| 118 | IMultiImage *_output_multi; |
| 119 | std::array<uint32_t, 3> _x_subsampling; |
| 120 | std::array<uint32_t, 3> _y_subsampling; |
| 121 | unsigned int _num_elems_processed_per_iteration; |
| 122 | bool _is_parallelizable; |
| 123 | }; |
Gian Marco Iodice | 356f643 | 2017-09-22 11:32:21 +0100 | [diff] [blame] | 124 | } // namespace arm_compute |
Anthony Barbier | 6ff3b19 | 2017-09-04 18:44:23 +0100 | [diff] [blame] | 125 | #endif /* __ARM_COMPUTE_NECHANNELCOMBINEKERNEL_H__ */ |