Scribd
Upload
40 views27 pages

Structure Texture Filling-In of Missing Image Blocks: - An Image Processing Technique

This document describes a Structure Texture Filling-In (STF) technique for filling in missing image blocks when images are transmitted over fading wireless channels. The STF technique classifies lost blocks as either structured or textured and uses different algorithms to reconstruct each type of block. Structured blocks are filled in using image inpainting, while textured blocks are reconstructed using texture synthesis from surrounding blocks. The STF technique aims to reconstruct lost image data without needing additional error correction packets.

Uploaded by

Mani Pavan
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
40 views27 pages

Structure Texture Filling-In of Missing Image Blocks: - An Image Processing Technique

This document describes a Structure Texture Filling-In (STF) technique for filling in missing image blocks when images are transmitted over fading wireless channels. The STF technique classifies lost blocks as either structured or textured and uses different algorithms to reconstruct each type of block. Structured blocks are filled in using image inpainting, while textured blocks are reconstructed using texture synthesis from surrounding blocks. The STF technique aims to reconstruct lost image data without needing additional error correction packets.

Uploaded by

Mani Pavan
Copyright
© Attribution Non-Commercial (BY-NC)
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 27

STF

Structure Texture Filling-In of Missing Image Blocks

-AN IMAGE PROCESSING TECHNIQUE

Abstract:
An Approach for filling-in blocks of Missing data in Wireless Image Transmission. When Images are transmitted over fading channels,the effects of noise can destroy entire blocks of the image. Here we aim to reconstruct the lost data using correlation between the lost block and its neighbors. If the lost block contained structure,it is reconstructed using an image In-painting algorithm. If the lost block is textured block Texture-Synthesis is used.

Introduction
Images are transmitted over the wireless channel block by block. Due to severe fading, we may loss an entire block,even several consecutive blocks of an image Average packet loss rate in a wireless environment is 3.6% and occurs in a busty fashion. In the worst case ,a whole line of image blocks might be lost. Even if a single block is lost,the remaining blocks in that line might be received without their correct average value.

Why should we use STF


We have two common techniques to make the transmission robust are FEC and ARQ. But here the drawback is FEC needs extra error correction packets to be transmitted. ARQ lowers data transmission rates and further increases network congestion. So we go for STF to satisfactorily reconstruct the lost blocks using available information surrounding them.

System Analysis
Existing System
Network congestion increases.
Packet loss ratio is also increases

Proposed System
Band width efficiency increases. Network congestion decreases

Objective of the Project


The objective of this project is to filling in of lossy blocks without any extra packets needed along with the original image. We can reconstruct the lossy blocks using neighbor pixels with this project

Overview
Three major steps are involved in this project. They are 1. Classification of lossy blocks in to structured or textured 2. Use texture synthesis for texture blocks 3. Use image inpainting for structured blocks

Block Classification
We must Classify the lost block into texture or structure. If gray value of a pixel or block is lost, then it is considered as structured block. Filling of those blocks is done with the neighbor pixel colors using Image In-Painting. If the entire pixel or block is lost then it is considered as textured blocks. Filling of those blocks is done using texture synthesis

Texture Synthesis
After the classification of block as texture, then Texture Synthesis is applied on the block. The missing block is then filled-in with the texture from its surrounding. Let the region to be filled be denoted by S. The lost block will now be filled, pixel by pixel, in a raster fashion.

Steps Involved:

Steps Involved
Consider a template which touches the first pixel of the lost block. Now calculate sum of the squares of 9 gray values in 3*3 template. Now consider 10 random 3*3 blocks near by the template. Calculate sum of the squares of gray values for each block. Calculate the differences between the sums of the square values of the 10 random blocks with the template. Now consider the block as a candidate block which contains lowest difference from the template. Now take the pixel, which is at the immediate right of the candidate block and fill it in the first pixel of the lost block. In this way all the pixels of structure blocks are filled with Texture Synthesis

Image Inpainting
Structure in an image can be an edge between two
regions or a deterministic change in color or gray value. When the block classification algorithm detected a structured block, this is restored using the digital inpainting procedure. In image inpainting, the lossy gray values of the pixels are filled by the neighbor pixels gray values..

Steps Involved
The general idea behind this image inpainting technique is to look at the contours of the image and try to extend them into the masked region in a continuous manner. This technique tries to extend these edges and the colors between them directly. Inpainting technique consists of several iterations of "inpainting". If the bottom and right pixels of a pixel are black then the black pixels are filled by the above pixel. In this way all the pixels are filled by image inpainting

Mean Filter
After the completion of image inpainting on image,
mean filter is occurred. Mean Filter takes the pixels filled by the image inpainting. Those pixels are filled with average gray values of the surrounding 9 pixels. Mean filter may be done in several iterations. Generally two iterations are occurred.

System Design
Use Case Diagram for STF Sequence Diagram for STF Class Diagram for STF

Use Case Diagram

Sequence Diagram

Class Diagram

Image Details
Pixels:
An image is a rectangular array of dots called pixels (picture elements) where the number of rows M and number of columns N of dots in an image are specified. At each row-column intersection (m, n) there is a pixel, or picture element. The point (m, n) is the location of the pixel, while the pixel value at that location is designated by p(m, n), or sometimes by f(m, n) or f(x, y).

Image Co-ordinates
The following figure shows the location (m, n), where 0 < m < M-1 and 0 < n < N-1.

Pixel Values
The pixel values in an image may be:
Grayscale (or) Color

Grayscale Pixel Values


The grayscale images are simpler and each pixel value corresponds to one byte. Grayscale usually has a range from 0 (no light intensity, or full black) to 255 (full light intensity, or full white), in integer steps Each of these takes one byte of storage in a computer or in a file.

Color Pixel Values


. The most common method is called true-color and uses one byte for each of red, green and blue. Thus a single pixel value requires 3 bytes of memory or disk storage. It is clear that more colors and more pixels are more costly in computer storage and time to send on the Internet.

Conclusion
In this project we have proposed a new technique for
the filling-in of missing blocks in transmission of compressed images. We have shown that as long as the features in the image are not completely lost, they can be satisfactorily reconstructed using image inpainting and texture synthesis algorithms. This eliminates the need for retransmission of lost blocks

Future Enhancements
This project can be extended as the filling of lossy blocks in remaining image formats except JPG and PGM images.

Queries ??

You might also like