Mech-DLK SDK (C++) 3.0.0
Mech-DLK SDK (C++) Reference Documentation
 
Loading...
Searching...
No Matches
mmind::dl Namespace Reference

Classes

struct  MBbox
 
struct  MBlobValue
 
struct  MContour
 Represents a contour with hierarchical structure and associated blob features. More...
 
struct  MMindImage
 Represents an image structure for inference and visualization. More...
 
class  MMindInferEngine
 Defines the inference engine for Mech-DLK model packages. More...
 
struct  MMindResult
 Represents the generic result data structure for algorithm modules. More...
 
struct  ModelEfficiencyParam
 
struct  MPoint
 
struct  MPointF
 
struct  MRotatedBbox
 Rotated bounding box for object detection. More...
 

Enumerations

enum class  BackendType : int32_t {
  Unknown = 0 , OnnxRT , TensorRT , OpenVINO ,
  OnnxRTC
}
 
enum class  DLAlgoType : int32_t {
  Unknown = 0 , Classification , ObjectDetection , InstanceSegmentation ,
  InstanceSegmentationV2 , DefectSegmentation , FastLocating , TextRecognition ,
  TextDetection , UnsupSegmentation , LargeSegModel , FGBGSegmentation ,
  PickAnything , RegisterClassification , SingleCharacterDetection , RegisterUnsupSeg ,
  RegisterCharacterDetection
}
 
enum class  QuickAIMode : int32_t { Registration = 0 , Inference }
 
enum class  PrecisionType : int32_t { FP16 , FP32 }
 
enum class  InferDeviceType { Unknown = 0 , CPU = 1 , GPUDefault = 1 << 1 , GPUOptimization = 1 << 2 }
 
enum class  ModelStatus : unsigned {
  Idle , Converting , ConvertFailed , Converted ,
  Loading , LoadFailed , Loaded
}
 
enum class  InferImageType { Depth = 0x0 , Intensity = 0x1 }
 

Functions

constexpr MPointF operator+ (const MPointF &p1, const MPointF &p2)
 
MPoint operator+ (const MPoint &p1, const MPoint &p2)
 
MBlobValue blobValueFromContour (const MContour &contour)
 Computes region feature values (area, width, height, aspect ratio, etc.) for a contour.
 

Enumeration Type Documentation

◆ BackendType

enum class mmind::dl::BackendType : int32_t
strong
Enumerator
Unknown 
OnnxRT 
TensorRT 
OpenVINO 
OnnxRTC 

Definition at line 23 of file type.h.

◆ DLAlgoType

enum class mmind::dl::DLAlgoType : int32_t
strong
Enumerator
Unknown 
Classification 
ObjectDetection 
InstanceSegmentation 
InstanceSegmentationV2 
DefectSegmentation 
FastLocating 
TextRecognition 
TextDetection 
UnsupSegmentation 
LargeSegModel 
FGBGSegmentation 
PickAnything 
RegisterClassification 
SingleCharacterDetection 
RegisterUnsupSeg 
RegisterCharacterDetection 

Definition at line 31 of file type.h.

31 : int32_t {
32 Unknown = 0,
49};
@ RegisterCharacterDetection
Definition type.h:48

◆ InferDeviceType

enum class mmind::dl::InferDeviceType
strong
Enumerator
Unknown 
CPU 

CPU mode. It is used when a GPU is not available.

GPUDefault 

GPU default mode.

GPUOptimization 

GPU optimization mode. In this mode, it takes 1-5 minutes to optimize the model.

Definition at line 55 of file type.h.

55 {
56 Unknown = 0,
57 CPU = 1,
58 GPUDefault = 1 << 1,
59 GPUOptimization = 1 << 2,
61};
@ CPU
CPU mode. It is used when a GPU is not available.
Definition type.h:57
@ GPUDefault
GPU default mode.
Definition type.h:58

◆ InferImageType

enum class mmind::dl::InferImageType
strong
Enumerator
Depth 
Intensity 

Definition at line 73 of file type.h.

73 {
74 Depth = 0x0,
75 Intensity = 0x1,
76};

◆ ModelStatus

enum class mmind::dl::ModelStatus : unsigned
strong
Enumerator
Idle 
Converting 
ConvertFailed 
Converted 
Loading 
LoadFailed 
Loaded 

Definition at line 63 of file type.h.

◆ PrecisionType

enum class mmind::dl::PrecisionType : int32_t
strong
Enumerator
FP16 
FP32 

Definition at line 53 of file type.h.

◆ QuickAIMode

enum class mmind::dl::QuickAIMode : int32_t
strong
Enumerator
Registration 
Inference 

Definition at line 51 of file type.h.

51: int32_t { Registration = 0, Inference };

Function Documentation

◆ blobValueFromContour()

MBlobValue mmind::dl::blobValueFromContour ( const MContour & contour)

Computes region feature values (area, width, height, aspect ratio, etc.) for a contour.

The returned MBlobValue can be used for subsequent filtering or other processing.

Parameters
[in]contourThe contour to analyze.
Returns
MBlobValue containing the computed region features. Returns a default-constructed MBlobValue if analysis yields no result.

◆ operator+() [1/2]

MPoint mmind::dl::operator+ ( const MPoint & p1,
const MPoint & p2 )
inline

Definition at line 27 of file MContour.h.

27{ return {p1.x + p2.x, p1.y + p2.y}; }

◆ operator+() [2/2]

MPointF mmind::dl::operator+ ( const MPointF & p1,
const MPointF & p2 )
constexpr

Definition at line 27 of file MBbox.h.

28{
29 return {p1.x + p2.x, p1.y + p2.y};
30}