hi
will u please send me the total data about dis topic-"DESIGN OF DEAF MANâ„¢S TELEPHONE USING DSP (Digital Signal Processor TMS320C31)".
my e mail is johnyb4u[at]gmail.com
i need it urgent
thanq
|
DESIGN OF “DEAF MAN’S TELEPHONE†USING DSP (Digital Signal Processor TMS320C31)
|
hi
will u please send me the total data about dis topic-"DESIGN OF DEAF MANâ„¢S TELEPHONE USING DSP (Digital Signal Processor TMS320C31)".
my e mail is johnyb4u[at]gmail.com
i need it urgent
thanq
11-12-2009, 06:04 PM
A DSP BASED VIDEOPHONE FOR THE HEARING IMPAIRED
USING VALLEDGE PROCESSED PICTURES (USING TMS32010 DSP)
Abstract
An image coding algorithm that enables hand sign language to be used for conversation between the deaf over a telephone network is described. Valledge detector(called so) produces images like cartoon at a resolution of 64 by 64. These images are compresed using the block oriented algorithm and transmitted at 9.6 kb/s using modems that are full duplex. TMS32010 DSP is used to implement the algorithm in real time
HARDWARE OVERVIEW
The entire feature extraction, data compression and framing operation is done by a single TMS32010 DSP linked to a framestore that has 4 pages of 64 by 64.
CODING ALGORITHM
1. Temporal Filtering and Block Change Detection
A new 8 bit input image is captured into one page of the framestore, and combined an 8 by 8 block at a time with the contents of another page holding the previous image, using a non-linear
temporal filter algorithm. This filter reduces camera noise by temporally averaging low-level signals, but avoids visible lag by allowing large temporal changes through unfiltered .
The valledge detector
For perceptually important features an operator which has a primary response to
luminance valleys and a secondary response to luminance edges is required hence the name 'valledge' detector. In addition to the valledge detection which creates black lines on a white background, a black-fill operationis performed which is done by simply thresholding the original pels. If a pel is found to have a luminance value below the pre-set threshold level then it is set to black, else the feature extraction routine is applied. The advantage of the black level filling is to render dark areas, for example the hair and the part of the clothing, to black.
Block oriented coding
A recursive algorithm is used whereby an N by N block (N a power of 2 ) is said to be codeable if it is all 0 ' s (black) or all 1 s (white), otherwise it is uncodeable and is split down into four sub-blocks.
The detailed report download is only available at the IEEE Xplore digital library:
http://ieeexplore.ieeexpls/abs_all.jsp?a...6817&tag=1
You may access it if your college has a subscription or you or a friend having membership.
Cheers!
USING VALLEDGE PROCESSED PICTURES (USING TMS32010 DSP)
Abstract
An image coding algorithm that enables hand sign language to be used for conversation between the deaf over a telephone network is described. Valledge detector(called so) produces images like cartoon at a resolution of 64 by 64. These images are compresed using the block oriented algorithm and transmitted at 9.6 kb/s using modems that are full duplex. TMS32010 DSP is used to implement the algorithm in real time
HARDWARE OVERVIEW
The entire feature extraction, data compression and framing operation is done by a single TMS32010 DSP linked to a framestore that has 4 pages of 64 by 64.
CODING ALGORITHM
1. Temporal Filtering and Block Change Detection
A new 8 bit input image is captured into one page of the framestore, and combined an 8 by 8 block at a time with the contents of another page holding the previous image, using a non-linear
temporal filter algorithm. This filter reduces camera noise by temporally averaging low-level signals, but avoids visible lag by allowing large temporal changes through unfiltered .
The valledge detector
For perceptually important features an operator which has a primary response to
luminance valleys and a secondary response to luminance edges is required hence the name 'valledge' detector. In addition to the valledge detection which creates black lines on a white background, a black-fill operationis performed which is done by simply thresholding the original pels. If a pel is found to have a luminance value below the pre-set threshold level then it is set to black, else the feature extraction routine is applied. The advantage of the black level filling is to render dark areas, for example the hair and the part of the clothing, to black.
Block oriented coding
A recursive algorithm is used whereby an N by N block (N a power of 2 ) is said to be codeable if it is all 0 ' s (black) or all 1 s (white), otherwise it is uncodeable and is split down into four sub-blocks.
The detailed report download is only available at the IEEE Xplore digital library:
http://ieeexplore.ieeexpls/abs_all.jsp?a...6817&tag=1
You may access it if your college has a subscription or you or a friend having membership.
Cheers!
|
« Next Oldest | Next Newest »
|
