NVIDIA DeepStream SDK API Reference

6.0.1 Release

 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
nvdspreprocess_meta.h
Go to the documentation of this file.
1 
38 #ifndef __NVDSPREPROCESS_META_H__
39 #define __NVDSPREPROCESS_META_H__
40 
41 #include <vector>
42 #include <string>
43 #include "nvbufsurface.h"
44 #include "nvds_roi_meta.h"
45 
50 typedef struct
51 {
54 
56  guint64 buffer_size;
57 
59  std::vector<int> tensor_shape;
60 
63 
65  std::string tensor_name;
66 
68  guint gpu_id;
69 
71  void *private_data;
72 
74 
79 typedef struct
80 {
82  std::vector<guint64> target_unique_ids;
83 
86 
87  /* list of roi vectors per batch */
88  std::vector<NvDsRoiMeta> roi_vector;
89 
91  void *private_data;
92 
94 
95 #endif /* __NVDSPREPROCESS_META_H__ */
void * private_data
pointer to buffer from convert pool
void * private_data
reserved for private use
NvBufSurface Interface
std::vector< guint64 > target_unique_ids
target unique ids for which meta is prepared
Copyright (c) 2021, NVIDIA CORPORATION.
std::vector< NvDsRoiMeta > roi_vector
void * raw_tensor_buffer
raw tensor buffer preprocessed for infer
NvDsDataType
Data type used for model in infer.
Definition: nvds_roi_meta.h:56
NvDsDataType data_type
model datatype for which tensor prepared
guint gpu_id
gpu-id on which tensor prepared
std::string tensor_name
to be same as model input layer name
guint64 buffer_size
size of raw tensor buffer
tensor meta containing prepared tensor and related info inside preprocess user meta which is attached...
NvDsPreProcessTensorMeta * tensor_meta
pointer to tensor meta
preprocess meta as a user meta which is attached at batch level
std::vector< int > tensor_shape
raw tensor buffer shape