1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
|
/*
* Copyright (c) 2017-2020 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.
*/
#ifndef ARM_COMPUTE_CLCANNYEDGEKERNEL_H
#define ARM_COMPUTE_CLCANNYEDGEKERNEL_H
#include "arm_compute/core/CL/ICLKernel.h"
namespace arm_compute
{
class ICLTensor;
/** OpenCL kernel to perform Gradient computation.
*/
class CLGradientKernel : public ICLKernel
{
public:
/** Constructor */
CLGradientKernel();
/** Prevent instances of this class from being copied (As this class contains pointers) */
CLGradientKernel(const CLGradientKernel &) = delete;
/** Prevent instances of this class from being copied (As this class contains pointers) */
CLGradientKernel &operator=(const CLGradientKernel &) = delete;
/** Initialise the kernel's sources, destinations and border mode.
*
* @note gx, gy and mag must all be the same size (either 16 or 32).
*
* @param[in] gx Source tensor - Gx component. Data types supported: S16/S32.
* @param[in] gy Source tensor - Gy component. Data types supported: Same as gx.
* @param[out] magnitude Destination tensor - Magnitude. Data types supported: U16/U32. Must match the pixel size of gx, gy.
* @param[out] phase Destination tensor - Quantized phase. Data types supported: U8.
* @param[in] norm_type Normalization type. if 1, L1-Norm otherwise L2-Norm.
*/
void configure(const ICLTensor *gx, const ICLTensor *gy, ICLTensor *magnitude, ICLTensor *phase, int32_t norm_type);
/** Initialise the kernel's sources, destinations and border mode.
*
* @note gx, gy and mag must all be the same size (either 16 or 32).
*
* @param[in] compile_context The compile context to be used.
* @param[in] gx Source tensor - Gx component. Data types supported: S16/S32.
* @param[in] gy Source tensor - Gy component. Data types supported: Same as gx.
* @param[out] magnitude Destination tensor - Magnitude. Data types supported: U16/U32. Must match the pixel size of gx, gy.
* @param[out] phase Destination tensor - Quantized phase. Data types supported: U8.
* @param[in] norm_type Normalization type. if 1, L1-Norm otherwise L2-Norm.
*/
void configure(const CLCompileContext &compile_context, const ICLTensor *gx, const ICLTensor *gy, ICLTensor *magnitude, ICLTensor *phase, int32_t norm_type);
// Inherited methods overridden:
void run(const Window &window, cl::CommandQueue &queue) override;
private:
const ICLTensor *_gx; /**< Source tensor - Gx component */
const ICLTensor *_gy; /**< Source tensor - Gy component */
ICLTensor *_magnitude; /**< Destination tensor - Magnitude */
ICLTensor *_phase; /**< Destination tensor - Quantized phase */
};
/** OpenCL kernel to perform Non-Maxima suppression for Canny Edge.
*
* @note This kernel is meant to be used alongside CannyEdge and performs a non-maxima suppression using magnitude and phase of input
* to characterize points as possible edges. The output buffer needs to be cleared before this kernel is executed.
*
* @note Hysteresis is computed in @ref CLEdgeTraceKernel
*/
class CLEdgeNonMaxSuppressionKernel : public ICLKernel
{
public:
/** Constructor */
CLEdgeNonMaxSuppressionKernel();
/** Prevent instances of this class from being copied (As this class contains pointers) */
CLEdgeNonMaxSuppressionKernel(const CLEdgeNonMaxSuppressionKernel &) = delete;
/** Prevent instances of this class from being copied (As this class contains pointers) */
CLEdgeNonMaxSuppressionKernel &operator=(const CLEdgeNonMaxSuppressionKernel &) = delete;
/** Initialise the kernel's sources, destination and border mode.
*
* @param[in] magnitude Source tensor - Magnitude. Data types supported: U16/U32.
* @param[in] phase Source tensor - Quantized phase. Data types supported: U8.
* @param[out] output Destination tensor. Data types supported: U16/U32.
* @param[in] lower_thr Lower threshold.
* @param[in] border_undefined True if the border mode is undefined. False if it's replicate or constant.
*/
void configure(const ICLTensor *magnitude, const ICLTensor *phase, ICLTensor *output, int32_t lower_thr, bool border_undefined);
/** Initialise the kernel's sources, destination and border mode.
*
* @param[in] compile_context The compile context to be used.
* @param[in] magnitude Source tensor - Magnitude. Data types supported: U16/U32.
* @param[in] phase Source tensor - Quantized phase. Data types supported: U8.
* @param[out] output Destination tensor. Data types supported: U16/U32.
* @param[in] lower_thr Lower threshold.
* @param[in] border_undefined True if the border mode is undefined. False if it's replicate or constant.
*/
void configure(const CLCompileContext &compile_context, const ICLTensor *magnitude, const ICLTensor *phase, ICLTensor *output, int32_t lower_thr, bool border_undefined);
// Inherited methods overridden:
void run(const Window &window, cl::CommandQueue &queue) override;
BorderSize border_size() const override;
private:
const ICLTensor *_magnitude; /**< Source tensor - Magnitude. */
const ICLTensor *_phase; /**< Source tensor - Quantized phase. */
ICLTensor *_output; /**< Destination tensor. */
};
/** OpenCL kernel to perform Edge tracing.
*/
class CLEdgeTraceKernel : public ICLKernel
{
public:
/** Constructor */
CLEdgeTraceKernel();
/** Prevent instances of this class from being copied (As this class contains pointers) */
CLEdgeTraceKernel(const CLEdgeTraceKernel &) = delete;
/** Prevent instances of this class from being copied (As this class contains pointers) */
CLEdgeTraceKernel &operator=(const CLEdgeTraceKernel &) = delete;
/** Initialise the kernel's source, destination and border mode.
*
* @param[in] input Source tensor. Data types supported: U16/U32.
* @param[out] output Destination tensor. Data types supported: U8.
* @param[in] upper_thr Upper threshold used for the hysteresis
* @param[in] lower_thr Lower threshold used for the hysteresis
* @param[in,out] visited Tensor for keeping the visited pixels. Data types supported: U32.
* Expected to be initialized to 0 before each run.
* @param[in,out] recorded Tensor for keeping the recorded pixels. Data types supported: U32
* Expected to be initialized to 0 before each run.
* @param[in,out] l1_stack Tensor with the L1 stack for each pixel. Data types supported: S32.
* Expected to be initialized to 0 before each run.
* @param[in,out] l1_stack_counter Tensor for counting the elements in the L1 stack of each pixel. Data types supported: U8.
* Expected to be initialized to 0 before each run.
*/
void configure(const ICLTensor *input, ICLTensor *output, int32_t upper_thr, int32_t lower_thr,
ICLTensor *visited, ICLTensor *recorded, ICLTensor *l1_stack, ICLTensor *l1_stack_counter);
/** Initialise the kernel's source, destination and border mode.
*
* @param[in] compile_context The compile context to be used.
* @param[in] input Source tensor. Data types supported: U16/U32.
* @param[out] output Destination tensor. Data types supported: U8.
* @param[in] upper_thr Upper threshold used for the hysteresis
* @param[in] lower_thr Lower threshold used for the hysteresis
* @param[in,out] visited Tensor for keeping the visited pixels. Data types supported: U32.
* Expected to be initialized to 0 before each run.
* @param[in,out] recorded Tensor for keeping the recorded pixels. Data types supported: U32
* Expected to be initialized to 0 before each run.
* @param[in,out] l1_stack Tensor with the L1 stack for each pixel. Data types supported: S32.
* Expected to be initialized to 0 before each run.
* @param[in,out] l1_stack_counter Tensor for counting the elements in the L1 stack of each pixel. Data types supported: U8.
* Expected to be initialized to 0 before each run.
*/
void configure(const CLCompileContext &compile_context, const ICLTensor *input, ICLTensor *output, int32_t upper_thr, int32_t lower_thr,
ICLTensor *visited, ICLTensor *recorded, ICLTensor *l1_stack, ICLTensor *l1_stack_counter);
// Inherited methods overridden:
void run(const Window &window, cl::CommandQueue &queue) override;
private:
const ICLTensor *_input; /**< Source tensor. */
ICLTensor *_output; /**< Destination tensor. */
int32_t _lower_thr; /**< Lower threshold used for the hysteresis. */
int32_t _upper_thr; /**< Upper threshold used for the hysteresis. */
ICLTensor *_visited; /**< Marks visited elements */
ICLTensor *_recorded; /**< Marks recorded elements */
ICLTensor *_l1_stack; /**< L1 hysteris stack */
ICLTensor *_l1_stack_counter; /**< L1 hysteris stack counter */
};
} // namespace arm_compute
#endif /* ARM_COMPUTE_CLCANNYEDGEKERNEL_H */
|
