V3.1.0 (05/2016)
SDK version v3.1.0.0: release notes
Main Changes:
- Tutorial: How to develop your own Input Video Stream Plugin?
- New Face Recognition Model : A vastly improved describer for doing face recognition. DescriberDnn
- Removed old FaceDNNDescriber model
API Changes:
- papillon::PapillonSDK : SetLicense() now returns papillon::PResult
- papillon::PColour3f : this class has been removed
- papillon::PDescriptor : GetDescriptor() now returns papillon::PResult instead of bool
- papillon::PResult : constructor is now explicit to avoid automatic building from EReturnCode
- papillon::PVersion : is now immutable (Set() has become a private method)
API News:
- papillon::PEnrollment : some new functions for easily generating a description from images and videos
- papillon::PapillonSDK : add GetVersion()
- papillon::PFile : add IsReadable(), IsWritable(), IsExecutable(), CheckExistsAndIsReadable()
- papillon::PFrame : add SetProperties()
- papillon::PString : add ToTextFile(), FromTextFile()
Bug Fixed:
- papillon::PByteArray : fix Zip(): compressionLevel was not taken into account
- papillon::PInputVideoStream : better error message for Open(): now checks if URI is a file and if the file exists and is readable
- papillon::POption : better error management (e.g. when equal sign is missing on a long name option)
Performance improvements:
- no changes
Documentation:
- Some improvements and fixes
- Improvements to face recognition documentation
- Caller graphs have been removed (useless)
Plugins:
- FFMpeg input video source plugin (on external GitHub) now supports RTSP streams and GO_TO_TIME property