tnn_yolov5.h
3.0 KB
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
//
// Created by DefTruth on 2021/11/6.
//
#ifndef LITE_AI_TOOLKIT_TNN_CV_TNN_YOLOV5_H
#define LITE_AI_TOOLKIT_TNN_CV_TNN_YOLOV5_H
#include "lite/tnn/core/tnn_core.h"
namespace tnncv
{
class LITE_EXPORTS TNNYoloV5 : public BasicTNNHandler
{
public:
explicit TNNYoloV5(const std::string &_proto_path,
const std::string &_model_path,
unsigned int _num_threads = 1); //
~TNNYoloV5() override = default;
private:
// nested classes
typedef struct
{
float r;
int dw;
int dh;
int new_unpad_w;
int new_unpad_h;
bool flag;
} YoloV5ScaleParams;
private:
// In TNN: x*scale + bias
std::vector<float> scale_vals = {1.0 / 255.f, 1.0 / 255.f, 1.0 / 255.f}; // RGB
std::vector<float> bias_vals = {0.f, 0.f, 0.f};
const char *class_names[80] = {
"person", "bicycle", "car", "motorcycle", "airplane", "bus", "train", "truck", "boat", "traffic light",
"fire hydrant", "stop sign", "parking meter", "bench", "bird", "cat", "dog", "horse", "sheep", "cow",
"elephant", "bear", "zebra", "giraffe", "backpack", "umbrella", "handbag", "tie", "suitcase", "frisbee",
"skis", "snowboard", "sports ball", "kite", "baseball bat", "baseball glove", "skateboard", "surfboard",
"tennis racket", "bottle", "wine glass", "cup", "fork", "knife", "spoon", "bowl", "banana", "apple",
"sandwich", "orange", "broccoli", "carrot", "hot dog", "pizza", "donut", "cake", "chair", "couch",
"potted plant", "bed", "dining table", "toilet", "tv", "laptop", "mouse", "remote", "keyboard",
"cell phone", "microwave", "oven", "toaster", "sink", "refrigerator", "book", "clock", "vase",
"scissors", "teddy bear", "hair drier", "toothbrush"
};
enum NMS
{
HARD = 0, BLEND = 1, OFFSET = 2
};
static constexpr const unsigned int max_nms = 30000;
private:
void transform(const cv::Mat &mat_rs) override; // without resize
void resize_unscale(const cv::Mat &mat,
cv::Mat &mat_rs,
int target_height,
int target_width,
YoloV5ScaleParams &scale_params);
void generate_bboxes(const YoloV5ScaleParams &scale_params,
std::vector<types::Boxf> &bbox_collection,
std::shared_ptr<tnn::Instance> &_instance,
float score_threshold, int img_height,
int img_width); // rescale & exclude
void nms(std::vector<types::Boxf> &input, std::vector<types::Boxf> &output,
float iou_threshold, unsigned int topk, unsigned int nms_type);
public:
void detect(const cv::Mat &mat, std::vector<types::Boxf> &detected_boxes,
float score_threshold = 0.25f, float iou_threshold = 0.45f,
unsigned int topk = 100, unsigned int nms_type = NMS::OFFSET);
};
}
#endif //LITE_AI_TOOLKIT_TNN_CV_TNN_YOLOV5_H