Face recognition sunlightface
Face recognition accuracy rate of 97.18%, 1 second to complete.
Can be complex, and remains true in the context of light work, and supports more people face in one scene on the detection, identification, validation. Wearing glasses and black can also detect and identify. Detects human faces from various perspectives.
And brings people nose, mouth, spectacles, face about the border location feature.
This software is limited to visible light (such as indoor light, fluorescent lamps, energy saving lamps, incandescent lamps) and require users to actively cooperate with the mining templates, actively cooperate with recognition. it is advisable to have USB camera
Unable to use several solutions:
1. better not download to run on desktop, COPY C,D to any directory.
2.Win2003 cannot be used, we recommend using WINXP or WIN7;
WIN7 compatibility mode is recommended, to run as an administrator. Official version can be run under WIN2003.
3. determine is not running in the Debug class or virtual machine software such as SOFTICE,VMWARE.
4. monitor resolution is best to 1280*768 per cent, 1024*768 may not be used. Can't display color quality 16-bit color, if necessary, is a 24-bit or 32-bit colors.
5. If sunlightface.exe if no response, can be slightly wait, try, and try again. Shut down other applications or programming software as much as possible.
6. If the face cannot be detected after the first start. Recommended that you exit the software before you run again. If that doesn't work, can try to change WINXP Windows styles "topic" into classic or XP interface.
7. If you installed a USB webcam, but no video appears. Try USB_DEVICE_ID.txt USB device file ID number instead of 1 or 2, and so give it a try.
Face detection performance indicator:
1. the face angle: pitch up 30 degrees down below 30 degrees, 30 degrees to the left, 30 degrees to the right, the plane rotated 60 degrees.
2. light range: sunlight, indoor light, dark, light,.
3. face colour range: the yellow race and the white race, Brown, black, Lady makeup.
4. other support: blurry photos, color photographs, faded photos, glasses and hats.
5. image size: Max 1280*1280, 320*240 recommended, type: BMP/JPG. Up from a checked out 32 faces in the image.
6. face size: minimum eye base 8 pixels, 200 pixels maximum eye margin, recommend the best eye from 40 pixels.
7. face detection accuracy: (1500 picture random testing): 99.5%, General error rate 0.01%, knowledge spill rate of 0.04%.
8. face detection speed: image size 320*240,CPU:P4 2.4GHZ, average speed <200ms/.
Try USB_DEVICE_ID.txt USB device file ID number instead of 1 or 2, and so give it a try.
1. face recognition mode 1C1 speed: 320*240,CPU:P4 2.4GHZ image size is less than 1 second, (total number has nothing to do with the library).
2. face recognition mode 1CN speed: image size 320*240,CPU:P4 2.4GHZ 500 is 1 template library selected, less than 1 second; 1 selected in the 1500 template library is less than 2 seconds.
3.2GB memory is the largest template capacity is 100,000 (more than 100,000 will be slow), support database ACCESS,SQLSERVER, recommended personal template: 5-10.
4. this version is proactively identifying, she asked to be identified by looking at pictures of her, the right attitude seriously.
You can wear glasses, but make sure the eye is clearly visible in the picture. (Few glasses against white, and don't see the eyes! ）
Sun face-recognition Development Kit (SDK) compatibility information
1. facial recognition SDK platform: Windows, the source code is written in VC, but not GDI and third-party library, it can be easier to move to LINUX.
2. facial recognition SDK interface: SunLightFace.DLL, standard Windows API calls, with detailed instructions.
3. face recognition SDK supported languages. VC,VB,Delphi,C#。
(Delphi function when it is invoked, parameters and send output to do the conversion, VC a DLL can be used to test Delphi show the arguments passed are correct. )
4. classification of face recognition SDK version: active identification of visible light version (that is described in this page), passive identification visible version, proactively identifying an infrared version.
5. the authors other than the SDK technical support for the SDK provide facial recognition software secondary development technology support.
6. (proactively identifying visible version) add CFL lamp just above and to the lens filter, and when coupled with low illumination high definition cameras, will achieve the best results.
File listTips: You can preview the content of files by clicking file names^_^
|P1.cpp||8.79 kB||14-02-12 15:35|
|P1.h||546.00 B||14-02-12 15:34|
|P2.cpp||6.27 kB||14-02-12 15:35|
|P2.h||597.00 B||14-02-12 15:34|
|P3.cpp||9.14 kB||14-02-12 15:35|
|P3.h||616.00 B||14-02-12 15:34|
|P4.cpp||6.13 kB||14-02-12 15:35|
|P4.h||543.00 B||14-02-12 15:34|
|P5.cpp||7.03 kB||14-02-12 15:35|
|P5.h||512.00 B||14-02-12 15:34|
|P6.cpp||400.00 B||14-02-12 15:35|
|P6.h||294.00 B||14-02-12 15:34|
|P7.cpp||396.00 B||14-02-12 15:35|
|P7.h||305.00 B||14-02-12 15:34|
|ReadMe.txt||2.85 kB||24-05-05 18:21|
|table1.bmp||156.24 kB||06-06-09 17:43|
|tab_sheet_pro.ico||21.12 kB||25-01-02 21:04|
|tab_sheet_pro.manifest||702.00 B||24-05-05 18:21|
|tab_sheet_pro.rc2||380.00 B||24-05-05 18:21|
|TRFFC02.ICO||1.05 kB||05-06-09 19:58|
|_desktop.ini||9.00 B||20-05-07 15:26|
|Resource.h||2.37 kB||14-02-12 15:34|
|stdafx.cpp||142.00 B||14-02-12 15:35|
|stdafx.h||1.86 kB||14-02-12 15:34|
|SunLightData.dll||523.50 kB||02-10-09 13:56|
|SunLightFace.dll||427.50 kB||02-10-09 13:52|
|SunlightFace.exe||488.00 kB||01-10-09 18:55|
|SunLightFace.h||16.91 kB||14-02-12 15:34|
|SunLightFace.lib||7.87 kB||10-10-09 16:07|
|tab_sheet_pro.aps||225.30 kB||10-10-09 17:19|
|tab_sheet_pro.cpp||1.40 kB||14-02-12 15:35|
|tab_sheet_pro.h||489.00 B||14-02-12 15:34|
|tab_sheet_pro.rc||15.37 kB||10-10-09 17:19|
|tab_sheet_pro.sln.old||926.00 B||24-05-05 18:21|
|tab_sheet_pro.vcproj||5.98 kB||01-10-09 18:55|
|tab_sheet_pro.vcproj.old||4.47 kB||24-05-05 18:35|
|tab_sheet_proDlg.cpp||8.80 kB||14-02-12 15:35|
|tab_sheet_proDlg.h||1.90 kB||14-02-12 15:34|
|VC_TEST.sln||915.00 B||02-06-05 11:11|
|VC_TEST.suo||13.00 kB||14-02-12 12:58|
|zhardware.cpp||38.20 kB||14-02-12 15:35|
|zhardware.h||11.81 kB||14-02-12 15:34|
|zstring.cpp||36.65 kB||14-02-12 15:35|
|zstring.h||12.43 kB||14-02-12 15:34|
|zstringEx.cpp||37.68 kB||14-02-12 15:35|
|zstringEx.h||11.89 kB||14-02-12 15:34|
|<res>||0.00 B||26-12-13 23:39|
|<VC_sample>||0.00 B||27-12-13 18:32|
(Add your comment, get 0.1 Point)
- Page 1
- Total 1