Professional Documents
Culture Documents
Introduction
Slide 2
Overview
Slide 3
Slide 4
Slide 5
Slide 6
Slide 7
Slide 8
Slide 9
Slide 10
Slide 11
Slide 12
Slide 14
Concept of Digital Watermarking (it is a copy right protection of ownership of any multimedia objects
(such as image, video, audio and text))
Slide 15
Slide 16
Slide 17
The encoder takes watermark data and original or host image as inputs to produce the watermarked
image. It is also encrypted by a key data with only known by the sender and receiver.
In decoding, the decoder needs the original image to extract the watermarked image from the
watermarked image
Slide 18
The main different between a conventional digital watermarking and reversible watermarking
Slide 19
We are focusing on the three basic and well known algorithms for reversible watermarking. They are
1. Difference Expansion (mainly used when image pixels are in uniform region)
2. Improved Rhombus Interpolation (can be used on image pixels even they are not in uniform
region )
3. Adaptive Feedback based Difference Expansion (It gives us the guarantee to provide the
highest payload with higher PSNR at all times.) highest payload means higher embedding
capacity and higher PSNR(Peak Signal to Noise Ration) means higher different between
original and watermarked image (Both are depends on the SSIM which should not excess 1)
Slide 20
The name of institutes project: The proposed Adaptive Feedback (AF) using DE Based Reversible
Watermarking (RW)
Slide 21
Flow chart for Adaptive Feedback based RIW using DE. We added some additional steps to add an
adaptive feedback with existing Difference Expansion algorithm for reversible Watermarking to provide
highest payload with higher PSNR at all times.
Slide 22
Specification of FPGA based Reversible Watermarking. by taking a look on the other existing FPGA
based RW algorithms , This proposed FPGA based architecture of RW algorithm should follow the
following specification to becomes the best architecture among them
Power- Overall system power-( 4.4V -5V)
Frequency Clock frequency-(> 4.32 MHz)
I/O stands for Input and output ports:- (should be =<61% i/o port of the FPGA ) less i/o means
less execution time required to complete overall process.
LUT Look up table-used to adjust the contrast characteristics of an image there are several
LUT made for a single FPGA to adjust several properties of Images. Among them 0.14% is used
for adjusting the contract characteristics of an image.
Throughput The amount of data processed per clock cycle (bits per Second) -55Mbps for 170
MHz frequency means 55Mbps data should be processed per each 170 MHz sine wave frequency
(clock frequency)
Critical Path- Maximum Clock Frequency- The maximum delay between any two sequential
elements in a design will determine the max clock speed-10.67ns (for single cycle).
Avg. SSIM (Structural Similarity Index Matrix):- the SSIM between original and watermarked
image should be less than 1 to provide that the data of original image is changed due to embed the
watermark data. The SSIM between original and decoded image should be 1 as our purpose to get
back the original image from the watermarked image without any lost.
Avg. Fun Out:-gives an idea of how efficiently routed or congested our design may be, and
probably how likely to meet our timing constraints.
Latency The time between data input and processed data output (clock cycle) 242 ns
Slide 24
Slide 25
Slide 26
Slide 27
Budget Outlay
Slide 28
Manpower Details
Slide 29
Slide 30