MatchDescriptorsKnnFlannLsh
Match descriptors (k nearest neighbors) using a FLANN based approach with LSH index.
Syntax
Definition:
HRESULT MatchDescriptorsKnnFlannLsh(
HRESULT hrPrev,
ITcVnImage* ipQueryDescriptors,
ITcVnImage* ipTrainDescriptors,
ITcVnContainer*& ipMatches,
ULONG nK,
ULONG nTableNumber = 12,
ULONG nKeySize = 20,
ULONG nMultiProbeLevel = 2,
ULONG nChecks = 32
)Parameters
|
Name |
Type |
Default |
Description |
|---|---|---|---|
|
hrPrev |
|
HRESULT indicating the result of previous operations (If SUCCEEDED(hrPrev) equals false, no operation is executed.) | |
|
ipQueryDescriptors |
|
Query descriptors (i.e. descriptors of a reference/template image) | |
|
ipTrainDescriptors |
|
Training descriptors (i.e. descriptors of a source/input image) | |
|
ipMatches |
|
Returns a container which is filled with the descriptor matches (ContainerType_Vector_Vector_TcVnDMatch; Non-zero interface pointers are reused.) | |
|
nK |
ULONG |
|
Number of required best train matches for each query descriptor (i.e. nk := 2) |
|
nTableNumber |
ULONG |
12 |
Number of tables |
|
nKeySize |
ULONG |
20 |
Key size |
|
nMultiProbeLevel |
ULONG |
2 |
Multi-probe level |
|
nChecks |
ULONG |
32 |
Maximum number of visited leafs when searching for neighbors |
Return valueRequired License
TC3 Vision Matching