Skip to content

Commit

Permalink
Add: nvidia object detection algo
Browse files Browse the repository at this point in the history
* Add: nvidia algo register

* Add: nvidia object detection algo

* Optim: model valid
  • Loading branch information
mjq2020 authored Mar 27, 2024
1 parent 8ef2f15 commit 920f645
Show file tree
Hide file tree
Showing 6 changed files with 374 additions and 7 deletions.
6 changes: 6 additions & 0 deletions core/algorithm/el_algorithm_delegate.cpp
Original file line number Diff line number Diff line change
Expand Up @@ -52,6 +52,9 @@ el_algorithm_type_t el_algorithm_type_from_engine(const Engine* engine) {
#endif
#ifdef _EL_ALGORITHM_PFLD_H_ // index 4
if (AlgorithmPFLD::is_model_valid(engine)) return EL_ALGO_TYPE_PFLD;
#endif
#ifdef _EL_ALGORITHM_NVIDIA_DET_H_
if (AlgorithmNvidiaDet::is_model_valid(engine)) return EL_ALGO_TYPE_NVIDIA_DET;
#endif
return EL_ALGO_TYPE_UNDEFINED;
}
Expand Down Expand Up @@ -103,6 +106,9 @@ AlgorithmDelegate::AlgorithmDelegate() {
#ifdef _EL_ALGORITHM_YOLO_V8_H_
_registered_algorithms.emplace_front(&AlgorithmYOLOV8::algorithm_info);
#endif
#ifdef _EL_ALGORITHM_NVIDIA_DET_H_
_registered_algorithms.emplace_front(&AlgorithmNvidiaDet::algorithm_info);
#endif
}

} // namespace edgelab
1 change: 1 addition & 0 deletions core/algorithm/el_algorithm_delegate.h
Original file line number Diff line number Diff line change
Expand Up @@ -36,6 +36,7 @@
#include "el_algorithm_yolo.h"
#include "el_algorithm_yolo_pose.h"
#include "el_algorithm_yolov8.h"
#include "el_algorithm_nvidia_det.h"

namespace edgelab {

Expand Down
232 changes: 232 additions & 0 deletions core/algorithm/el_algorithm_nvidia_det.cpp
Original file line number Diff line number Diff line change
@@ -0,0 +1,232 @@
/*
* The MIT License (MIT)
*
* Copyright (c) 2024 Seeed Technology Co.,Ltd
*
* 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 "el_algorithm_nvidia_det.h"

#include <cmath>
#include <type_traits>

#include "core/el_common.h"
#include "core/el_debug.h"
#include "core/utils/el_cv.h"
#include "core/utils/el_nms.h"

namespace edgelab {

AlgorithmNvidiaDet::InfoType AlgorithmNvidiaDet::algorithm_info{types::el_algorithm_nvidia_det_config_t::info};

AlgorithmNvidiaDet::AlgorithmNvidiaDet(EngineType* engine, ScoreType score_threshold, IoUType iou_threshold)
: Algorithm(engine, AlgorithmNvidiaDet::algorithm_info),
_w_scale(1.f),
_h_scale(1.f),
_score_threshold(score_threshold),
_iou_threshold(iou_threshold) {
init();
}

AlgorithmNvidiaDet::AlgorithmNvidiaDet(EngineType* engine, const ConfigType& config)
: Algorithm(engine, config.info),
_w_scale(1.f),
_h_scale(1.f),
_score_threshold(config.score_threshold),
_iou_threshold(config.iou_threshold) {
init();
}

AlgorithmNvidiaDet::~AlgorithmNvidiaDet() {
_results.clear();
this->__p_engine = nullptr;
}

bool AlgorithmNvidiaDet::is_model_valid(const EngineType* engine) {
const auto& input_shape{engine->get_input_shape(0)};
if (input_shape.size != 4 || // B, W, H, C
input_shape.dims[0] != 1 || // B = 1
input_shape.dims[1] ^ input_shape.dims[2] || // W = H
input_shape.dims[1] < 32 || // W, H >= 32
input_shape.dims[1] % 16 || // W or H is multiply of 32
(input_shape.dims[3] != 3 && // C = RGB or Gray
input_shape.dims[3] != 1))
return false;
auto output_shape0 = engine->get_output_shape(0);
auto output_shape1 = engine->get_output_shape(1);

if (output_shape0.size < 4 || output_shape1.size < 4) return false;

if (input_shape.dims[1] / 16 != output_shape0.dims[1] || input_shape.dims[2] / 16 != output_shape0.dims[2] ||
input_shape.dims[1] / 16 != output_shape1.dims[1] || input_shape.dims[2] / 16 != output_shape1.dims[2])
return false;

return true;
}

inline void AlgorithmNvidiaDet::init() {
EL_ASSERT(is_model_valid(this->__p_engine));
EL_ASSERT(_score_threshold.is_lock_free());
EL_ASSERT(_iou_threshold.is_lock_free());

_input_img.data = static_cast<decltype(ImageType::data)>(this->__p_engine->get_input(0));
_input_img.width = static_cast<decltype(ImageType::width)>(this->__input_shape.dims[1]),
_input_img.height = static_cast<decltype(ImageType::height)>(this->__input_shape.dims[2]),
_input_img.size =
static_cast<decltype(ImageType::size)>(_input_img.width * _input_img.height * this->__input_shape.dims[3]);
_input_img.format = EL_PIXEL_FORMAT_UNKNOWN;
_input_img.rotate = EL_PIXEL_ROTATE_0;
if (this->__input_shape.dims[3] == 3) {
_input_img.format = EL_PIXEL_FORMAT_RGB888;
} else if (this->__input_shape.dims[3] == 1) {
_input_img.format = EL_PIXEL_FORMAT_GRAYSCALE;
}
EL_ASSERT(_input_img.format != EL_PIXEL_FORMAT_UNKNOWN);
EL_ASSERT(_input_img.rotate != EL_PIXEL_ROTATE_UNKNOWN);
}

el_err_code_t AlgorithmNvidiaDet::run(ImageType* input) {
_w_scale = static_cast<float>(input->width) / static_cast<float>(_input_img.width);
_h_scale = static_cast<float>(input->height) / static_cast<float>(_input_img.height);

// TODO: image type conversion before underlying_run, because underlying_run doing a type erasure
return underlying_run(input);
};

el_err_code_t AlgorithmNvidiaDet::preprocess() {
auto* i_img{static_cast<ImageType*>(this->__p_input)};

// convert image
el_img_convert(i_img, &_input_img);

auto size{_input_img.size};
for (decltype(ImageType::size) i{0}; i < size; ++i) {
_input_img.data[i] -= 128;
}

return EL_OK;
}

float min(float a, float b) {
if (a < b)
return a;
else
return b;
}

float max(float a, float b) {
if (a > b)
return a;
else
return b;
}

el_err_code_t AlgorithmNvidiaDet::postprocess() {
_results.clear();
el_shape_t __output_shape0;
el_shape_t __output_shape1;
// get output
auto* data0{static_cast<_Float32*>(this->__p_engine->get_output(0))};
auto* data1{static_cast<_Float32*>(this->__p_engine->get_output(1))};
__output_shape0 = this->__p_engine->get_output_shape(0);
__output_shape1 = this->__p_engine->get_output_shape(1);

auto* bboxs = __output_shape0.dims[3] > __output_shape1.dims[3] ? data0 : data1;
auto* conf = __output_shape0.dims[3] > __output_shape1.dims[3] ? data1 : data0;

this->_conf_shape = __output_shape0.dims[3] > __output_shape1.dims[3] ? __output_shape1 : __output_shape0;
this->_bboxes_shape = __output_shape0.dims[3] > __output_shape1.dims[3] ? __output_shape0 : __output_shape1;

auto B = this->_conf_shape.dims[0];
auto H = this->_conf_shape.dims[1];
auto W = this->_conf_shape.dims[2];
auto BboxsCount = this->_conf_shape.dims[3];
auto C = BboxsCount * 4;

for (int h = 0; h < H; h++) {
for (int w = 0; w < W; w++) {
for (int j = 0; j < BboxsCount; j++) {
if (conf[h * (W * BboxsCount) + w * BboxsCount + j] > 0.2) {
BoxType box{
.x = 0,
.y = 0,
.w = 0,
.h = 0,
.score = 0,
.target = 0,
};
box.x = max(w * this->stride + this->offset - bboxs[h * (W * C) + w * C + j * 4] * this->scale, 0) *
this->_w_scale;
box.y =
max(h * this->stride + this->offset - bboxs[h * (W * C) + w * C + j * 4 + 1] * this->scale, 0) *
this->_h_scale;
box.w = (min(w * this->stride + this->offset + bboxs[h * (W * C) + w * C + j * 4 + 2] * this->scale,
W * this->stride)) *
this->_w_scale -
box.x - 1;
box.h = (min(h * this->stride + this->offset + bboxs[h * (W * C) + w * C + j * 4 + 3] * this->scale,
H * this->stride)) *
this->_h_scale -
box.y - 1;
box.x = box.x + box.w / 2;
box.y = box.y + box.h / 2;

box.score = (int)(conf[h * (W * BboxsCount) + w * BboxsCount + j] * 200);
box.target = j;
_results.emplace_front(std::move(box));
}
}
}
}

ScoreType score_threshold{get_score_threshold()};
IoUType iou_threshold{get_iou_threshold()};

el_nms(_results, iou_threshold, score_threshold, false, true);

// _results.sort([](const BoxType& a, const BoxType& b) { return a.x < b.x; });

return EL_OK;
}

const std::forward_list<AlgorithmNvidiaDet::BoxType>& AlgorithmNvidiaDet::get_results() const { return _results; }

void AlgorithmNvidiaDet::set_score_threshold(ScoreType threshold) { _score_threshold.store(threshold); }

AlgorithmNvidiaDet::ScoreType AlgorithmNvidiaDet::get_score_threshold() const { return _score_threshold.load(); }

void AlgorithmNvidiaDet::set_iou_threshold(IoUType threshold) { _iou_threshold.store(threshold); }

AlgorithmNvidiaDet::IoUType AlgorithmNvidiaDet::get_iou_threshold() const { return _iou_threshold.load(); }

void AlgorithmNvidiaDet::set_algorithm_config(const ConfigType& config) {
set_score_threshold(config.score_threshold);
set_iou_threshold(config.iou_threshold);
}

AlgorithmNvidiaDet::ConfigType AlgorithmNvidiaDet::get_algorithm_config() const {
ConfigType config;
config.score_threshold = get_score_threshold();
config.iou_threshold = get_iou_threshold();
return config;
}

} // namespace edgelab
115 changes: 115 additions & 0 deletions core/algorithm/el_algorithm_nvidia_det.h
Original file line number Diff line number Diff line change
@@ -0,0 +1,115 @@
/*
* The MIT License (MIT)
*
* Copyright (c) 2024 Seeed Technology Co.,Ltd
*
* 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 _EL_ALGORITHM_NVIDIA_DET_H_
#define _EL_ALGORITHM_NVIDIA_DET_H_

#include <atomic>
#include <cstdint>
#include <forward_list>

#include "core/el_types.h"
#include "el_algorithm_base.h"

namespace edgelab {

using namespace edgelab::base;
using namespace edgelab::types;

namespace types {

// we're not using inheritance since it not standard layout
struct el_algorithm_nvidia_det_config_t {
static constexpr el_algorithm_info_t info{
.type = EL_ALGO_TYPE_NVIDIA_DET, .categroy = EL_ALGO_CAT_DET, .input_from = EL_SENSOR_TYPE_CAM};
uint8_t score_threshold = 50;
uint8_t iou_threshold = 45;
};

} // namespace types

class AlgorithmNvidiaDet final : public Algorithm {
public:
using ImageType = el_img_t;
using BoxType = el_box_t;
using ConfigType = el_algorithm_nvidia_det_config_t;
using ScoreType = decltype(el_algorithm_nvidia_det_config_t::score_threshold);
using IoUType = decltype(el_algorithm_nvidia_det_config_t::iou_threshold);

static InfoType algorithm_info;

AlgorithmNvidiaDet(EngineType* engine, ScoreType score_threshold = 50, IoUType iou_threshold = 45);
AlgorithmNvidiaDet(EngineType* engine, const ConfigType& config);
~AlgorithmNvidiaDet();

static bool is_model_valid(const EngineType* engine);

el_err_code_t run(ImageType* input);
const std::forward_list<BoxType>& get_results() const;

void set_score_threshold(ScoreType threshold);
ScoreType get_score_threshold() const;

void set_iou_threshold(IoUType threshold);
IoUType get_iou_threshold() const;

void set_algorithm_config(const ConfigType& config);
ConfigType get_algorithm_config() const;

protected:
inline void init();

el_err_code_t preprocess() override;
el_err_code_t postprocess() override;

private:
enum {
INDEX_X = 0,
INDEX_Y = 1,
INDEX_W = 2,
INDEX_H = 3,
INDEX_T = 4,
};

ImageType _input_img;
float _w_scale;
float _h_scale;

int8_t stride = 16;
int8_t scale = 35;
float offset = 0.5;

el_shape_t _conf_shape;
el_shape_t _bboxes_shape;

std::atomic<ScoreType> _score_threshold;
std::atomic<IoUType> _iou_threshold;

std::forward_list<BoxType> _results;
};

} // namespace edgelab

#endif
15 changes: 8 additions & 7 deletions core/el_types.h
Original file line number Diff line number Diff line change
Expand Up @@ -167,13 +167,14 @@ typedef int el_model_format_v;
* @brief Algorithm Types
*/
typedef enum {
EL_ALGO_TYPE_UNDEFINED = 0u,
EL_ALGO_TYPE_FOMO = 1u,
EL_ALGO_TYPE_PFLD = 2u,
EL_ALGO_TYPE_YOLO = 3u,
EL_ALGO_TYPE_IMCLS = 4u,
EL_ALGO_TYPE_YOLO_POSE = 5u,
EL_ALGO_TYPE_YOLO_V8 = 6u,
EL_ALGO_TYPE_UNDEFINED = 0u,
EL_ALGO_TYPE_FOMO = 1u,
EL_ALGO_TYPE_PFLD = 2u,
EL_ALGO_TYPE_YOLO = 3u,
EL_ALGO_TYPE_IMCLS = 4u,
EL_ALGO_TYPE_YOLO_POSE = 5u,
EL_ALGO_TYPE_YOLO_V8 = 6u,
EL_ALGO_TYPE_NVIDIA_DET = 7u,
} el_algorithm_type_t;

/**
Expand Down
Loading

0 comments on commit 920f645

Please sign in to comment.