1.3.0.7

Bug Fix

• Age/Gender/Emotion classification does not work on Linux/OSX (#208)
• CropFace does not work for Greyscale image (#207)

1.3.0.6

New Features

• Support to estimate human face emotion (#194)

1.3.0.5

Bug Fix

• yield return occurs performance issue (#184)

1.3.0.4

New Feature

• Support CUDA 11.2 (#160)
• Add Image.ToBitmap method (#161)

1.3.0.3

New Feature

• Support to specify Model (#127)
• Support CUDA 11.0 and CUDA 11.1 (#144)
• Add property to get detected face confidence (#145)
• Add new overload for Create method (#148)

Fix

• Check location argument (#128)
• Disable Dropout for age/gender inference (#153)

1.3.0.2

New Feature

• Add FaceEncoding.GetRawEncoding method

Fix

• All 1.3.0.1 package depends on DlibDotNet cpu package (#115)

1.3.0.1

New Feature

• Add Head pose estimation
• Add Eye Blink Detector
• Add FaceRecognition.LoadImage(Bitmap bitmap)
• Add FaceRecognition.LoadImage(IntPtr, int, int, int, Mode)

Breaking

• Abolish FaceRecognition.LoadImage(byte[], int, int, int)
  • Please use LoadImage(byte[], int, int, int, Mode) instead of this
• Replace accessibility of AgeEstimator.RawPredict and AgeEstimator.RawPredictProbability with protected
• Replace accessibility of GenderEstimator.RawPredict and GenderEstimator.RawPredictProbability with protected

1.3.0.0

feat

• Add FaceRecognition.CropFaces method
• Add Image.Save method
• Add CustomFaceLandmarkDetection, CustomAgeEstimator and CustomGenderEstimator to FaceRecognition
  • You can use own custom data model files when you want to change it

Breaking

• Abolished PredictorModel.Helen
  • Use PredictorModel.Custom and set CustomFaceLandmarkDetection

1.2.3.14

feat

• Support Age and Gender classification

1.2.3.13

feat

• MKL and CUDA101