|
Minos (MinosCVC.dll) 14.1
|
|
|
Minos (MinosCVC.dll) 14.1
|
Data Structures | |
| struct | TFilterDef |
| Defines a linear convolution filter kernel (cXXYY) for use with FilterUser2x2, FilterUser3x3 and FilterUser5x5. More... | |
| struct | TLearnControlStructure |
| Parameter collection for the generation of Minos classifiers. More... | |
Typedefs | |
| typedef void * | CLF |
| Handle to a Minos classifier. More... | |
| typedef void * | MTS |
| Handle to a Minos training set. More... | |
| typedef void * | MTSIMAGE |
| Handle to a Minos training set image. More... | |
| typedef void * | MTSINSTANCE |
| Handle to a Minos training set instance. More... | |
| typedef void * | MTSMODEL |
| Handle to a Minos training set model. More... | |
| typedef void * | RESULTS |
| Handle to a list of Minos search results. More... | |
| typedef cvbbool_t(* | TFCallbackScan) (void *PrivateData, cvbval_t LinesToDo) |
| Callback for OptCorrelation and OptCorrelations. More... | |
| typedef cvbbool_t(* | TFCheckProgress) (void *PrivateData, MTSIMAGE CurImage, MTSMODEL CurModel, cvbval_t LinesToDo) |
| Callback signature that enables progress information during a Minos consistency check. More... | |
| typedef cvbbool_t(* | TFLearn) (void *PrivateData, MTSINSTANCE Instance, cvbval_t Total, cvbval_t Correct, cvbval_t Current) |
| Callback signature that enables progress information during the generation of a Minos classifier. More... | |
Enumerations | |
| enum | QualityMeasureMethod { Quality_Unnormalized = 0 , Quality_ReturnCorrelation = 1 } |
| Quality calculation methods for Minos classifiers. More... | |
| enum | TMinosCharacterEncoding { TMCE_CurrentCodePage = 0 , TMCE_UnicodeUTF8 = 1 } |
| Character encoding methods available when saving training sets or classifiers. More... | |
| typedef void* CLF |
Handle to a Minos classifier.
A Minos classifier is a reference counted object. Handle validity may be verified with the function IsCLF and it may be used with the CVCImg.dll functions ReleaseObject and ShareObject.
| typedef void* MTS |
Handle to a Minos training set.
A Minos training set is a reference counted object. Handle validity may be verified with the function IsMTS and it may be used with the CVCImg.dll functions ReleaseObject and ShareObject.
| typedef void* MTSIMAGE |
Handle to a Minos training set image.
A Minos training set image is a reference counted object. Handle validity may be verified with the function IsMTSImage and it may be used with the CVCImg.dll functions ReleaseObject and ShareObject.
MTSIMAGE handle is not an image like the ones referred to by an IMG handle! The MTSIMAGE is a container that combines an IMG with a list of the instances that have been marked in the image and a reference to owning training set. To retrieve the IMG from an MTSIMAGE use the function GetImageFromImage. | typedef void* MTSINSTANCE |
Handle to a Minos training set instance.
A Minos model is a reference counted object. Handle validity may be verified with the function IsMTSInstance and it may be used with the CVCImg.dll functions ReleaseObject and ShareObject.
| typedef void* MTSMODEL |
Handle to a Minos training set model.
A Minos model is a reference counted object. Handle validity may be verified with the function IsMTSModel and it may be used with the CVCImg.dll functions ReleaseObject and ShareObject.
| typedef void* RESULTS |
Handle to a list of Minos search results.
Such lists may be created by the functions SearchAll, ReadCharacterList and OptCorrelations. Minos result lists are reference counted objects. Handle validity may be verified with the function IsResults and they may be used with the CVCImg.dll functions ReleaseObject and ShareObject.
| typedef cvbbool_t(* TFCallbackScan) (void *PrivateData, cvbval_t LinesToDo) |
Callback for OptCorrelation and OptCorrelations.
This function will be called by OptCorrelation and OptCorrelations for every scan line in the area of interest that has been processed.
| [in] | PrivateData | Pointer to the user's private data. |
| [in] | LinesToDo | Remaining number of scan lines. |
TRUE to continue the operation that invoked the callback, and FALSE to interrupt it. | typedef cvbbool_t(* TFCheckProgress) (void *PrivateData, MTSIMAGE CurImage, MTSMODEL CurModel, cvbval_t LinesToDo) |
Callback signature that enables progress information during a Minos consistency check.
The function will be called by MTSConsistencyCheck, MTSImageCheck, MTSModelCheck and MTSModelImageCheck for every scan line of every image that is being processed. It provides the callee with information about the check's progress and an opportunity to interrupt the calling operation.
| [in] | PrivateData | Pointer to the user's private data. |
| [in] | CurImage | Handle to the currently processed MTS image. |
| [in] | CurModel | Handle to the currently processed MTS model. |
| [in] | LinesToDo | Remaining number of scan lines. |
TRUE to continue the operation that invoked the callback, and FALSE to interrupt it. | typedef cvbbool_t(* TFLearn) (void *PrivateData, MTSINSTANCE Instance, cvbval_t Total, cvbval_t Correct, cvbval_t Current) |
Callback signature that enables progress information during the generation of a Minos classifier.
This callback function will be called by LearnCLFFromMTS and LearnCLFFromMTSW after every instance that has been processed. It provides the callee with information about the learning progress and an opportunity to interrupt the learning process.
| [in] | PrivateData | Pointer to the user's private data. |
| [in] | Instance | Handle to the training set's most recently processed instance. |
| [in] | Total | Total number of instances in the training set. |
| [in] | Correct | Number of instances which so far been correctly recognized based on the preliminary classifier. Or, to put it differently, whenever the addition of new features wasn't necessary to correctly identify the next instance this counter is incremented. |
| [in] | Current | Index of the currently processed instance. |
TRUE to continue the operation that invoked the callback, and FALSE to interrupt it. | enum QualityMeasureMethod |
Quality calculation methods for Minos classifiers.
Minos classifiers support different ways of calculating the quality of search results. Select the approach a classifier should use by setting the correct value with the SetCLFQualityType function.
| Enumerator | |
|---|---|
| Quality_Unnormalized | Unnormalized quality feedback. Unnormalized quality computation is calculated as the average feature contrast over all the model's features but based on the image data of the object that has been found. As such, the result is in the range [0...255] with higher values indicating a higher quality. Note that with unnormalized result qualities the comparison between qualities from results that refer to different models is not encouraged as it may easily be misleading. If you are aiming for globally comparable results use Quality_ReturnCorrelation instead. |
| Quality_ReturnCorrelation | Normalized quality feedback. Quality is calculated as a correlation over all the features of a model versus the corresponding pixels in the image where the object has been found. This calculation method yield globally comparable results from the range [0...1] at only slightly higher processing time costs than Quality_Unnormalized. |
Character encoding methods available when saving training sets or classifiers.
Since Common Vision Blox 13.00.000 Minos supports Unicode by storing all the strings in serialized classifiers or training sets as UTF8 encoded strings. These files are still readable with previous versions of Minos, but all the strings will have a byte order map (basically three very weird special characters) prepended.
To avoid this and give users a possibility to store classifiers and training sets in a backward-compatible way the functions WriteCLFFileEx, WriteCLFFileExW, WriteMTSFileEx and WriteMTSFileExW have been introduced. These accept a parameter of type TMinosCharacterEncoding with which a backward compatible encoding may be selected.