Hi community member,
Please provide us the detail information(Input size, throughput, processing time and so on.)of Face detection demonstration by using C6748LCDK.
Best regards.
Kaka
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Hi,
Thanks for your post.
By default, the C6748 LCDK comes with the demo flashed in it. BIOS C6 Software development kit is FREE to download which contains the demos. You can download it from
http://software-dl.ti.com/dsps/dsps_public_sw/c6000/web/bios_c6sdk/latest/index_FDS.html.
After installation, you will find the face detect CCS project at "C:\Program Files\Texas Instruments\c6sdk_02_00_00_00\demos\facedetect\build\cgt_ccs\lcdk". By following the below wiki, you can run the demos.
The kit comes with a quick start guide that describes how the board need to setup the board for the demo. Also, the demo requires an external D1 camera to be connected to the board using a composite cable.
Here is a wiki link that shows the same :
You can also see a video of the demo on that wiki page
However OpenCV based implementation of haardetectobject can be used for doing face detect. Refer OpenCV for the algorithm. The c6accel release will have this function running on the DSP.
There is face detect demo available for internal BIOS5 release on C6748 DSP only device.
http://software-dl.ti.com/dsps/dsps_public_sw/c6000/web/bios_c6sdk/latest/index_FDS.html
Thanks & regards,
Sivaraj K
------------------------------------------------------------------------------------------------------
Please click the Verify Answer button on this post if it answers your question
-------------------------------------------------------------------------------------------------------
Hi Sivaraj,
Just additional questions. I saw this video below. Does the use of OMAP for face detection much better than using a Beaglebone Black for face detection?
- kel
Hi,
Thanks for your update.
I am not sure about the performance of facedetect algorithm used in Beaglebone black but with respect to C6748 device, OpenCV based implementation of haardetectobject facedeteact algorithm used is very effective and powerful algorithm which performs better in quality and throughput wise.
Thanks & regards,
Sivaraj K
------------------------------------------------------------------------------------------------------
Please click the Verify Answer button on this post if it answers your question
-------------------------------------------------------------------------------------------------------
Hi Sivaraj
Thank you for your detail information about face detection demonstration.
Would you please provide the specification/performance of this demonstration?
I would like to know the following information.
- Max Input video size (D1 only?)
- Delay time until input image analysis
If you have any questions, please let me know.
Best regards.
Kaka
Hi,
Thanks for your update.
Yes, it is for composite input D1 camera(640x480 resolution),
May be, please refer the below E2E post and the source code is attached (6758.OMAPL138_Image_sensor_test.zip) for the ARM and DSP interface with the image sensor here:
http://e2e.ti.com/support/dsp/omap_applications_processors/f/42/p/200759/716448.aspx#716448
Thanks & regards,
Sivaraj K