Image Compression Using Clustering Techniques

2013

This paper deals mainly with the image compression algorithms and presents a new color space normalization (CSN) technique for enhancing the discriminating power of color space along with the principal component analysis (PCA) which enables compression of colour images. Context-based modeling is an important step in image compression. We used optimized clustering algorithms to compress images effectively by making use of a large number of image contexts by separating a finite unlabeled data set into a finite and discrete set of natural, hidden data structures, rather than provide an accurate characterization of unobserved samples generated from the same probability distribution. Since images contain large number of varying density regions, we used an optimized density based algorithm from a pool. PCA is used to express the large 1-D vector of pixels constructed from 2-D color image into the compact principal components of the feature space. Each image may be represented as a weighte...

There is a correlation between pixels in each image so that each pixel value of adjacent pixels can be guessed. By removing these dependencies can be compressed images. Our goal is to reduce the amount of compressed image data needed to display the digital images and therefore reduce the cost of transmission and storage. Compression has a key role in many important applications. These applications include image database, transmission of images, remote sensing, medical imaging, military and space equipment remote control and so on. In addition to the compression, image coding, there's talk. That after quantization matrix should be coded range of conversions. In reconstruction after decoding to achieve our desired image obtained with the difference that the picture is far less than the original image. What we've done in this thesis using a fractal method utilizes a Kohonen neural networks and clustering to increase the compression ratio and reduction coding and decoding the image. We have implemented three methods based on fractal coding. The first method is simple fractal coding. In the second method to create the codebook of multiple tree fractal coding is used. In the second method of vector quantization LBG algorithm for Kohonen neural network-based clustering algorithm and code book for coding image is used. Results in the second method show faster encoding. The method is simple fractal compression rate is higher than other methods.

Multimedia data compression is a challenging situation for compression technique, due to the possibility of loss of data as well as it require large amount of storage place. The minimization of storage place and proper transmission of these data need compression. In this dissertation we proposed a block based DWT image compression technique using genetic algorithm and HCC code matrix. The HCC code matrix compressed into two different set redundant and non-redundant which generate similar pattern of block coefficient. The similar block coefficient generated by particle of swarm optimization. The process of particle of swarm optimization is select for the optimal block of DWT transform function. For the experimental purpose we used some standard image such as Lena, Barbara and cameraman image. The size of resolution of this image is 256*256. The source of image is Google.

2020

Demand for communication of multimedia data through the telecommunications network and Internet is growing exponentially. Significant portion of the multimedia data consists of images and they occupy the major portion of the communication bandwidth. The digitization of medical image information is a prime concern for the medical community. In this scenario, compression is an essential component for creating file sizes of manageable and transmittable dimensions. Image compression is nothing but decreasing the size of bytes of the file it can allows the user to store a huge amount of information within a static memory. This paper discusses an innovative proposed compression process to do lossy image compression by using clustering techniques such as K-Means, Fuzzy C-Means and Density Based Spatial Clustering of Applications with Noise (DBSCAN). The main objective is to compress the image where it is compressed to a reasonable size as well as preserving the quality of an image. The met...

In this paper, we propose two methods to construct the initial codebook for K-means clustering based on covariance and spectral decomposition. Experimental results with standard images show that the proposed methods produces better quality reconstructed images measured in terms of Peak Signal to Noise Ratio (PSNR).

Clustering is a unsupervised learning technique. This paper presents a clustering based technique that may be applied to Image compression. The proposed technique clusters all the pixels into predetermined number of groups and produces a representative color for each group. Finally for each pixel only clusters number is stored during compression. This technique can be obtained in machine learning which one of the best methods for clustering is. The k-medoids algorithm is a clustering algorithm related to the K-means algorithm and the medoidshift algorithm.

This paper proposes an image compression method using k-means learned features with a focus on compression ratio. Our method is a combination of lossy and lossless techniques. The lossy part splits the image into patches which are discretly encoded using a big, offline trained dictionary. The lossless part increases the entropy of the encoding by taking advantage of statistical dependencies between adjacent patches. Our method achieves better compression ratios than other state-of-the-art image compression algorithms with reasonable reconstruction errors and runtimes.

International Journal for Research in Applied Science & Engineering Technology (IJRASET), 2022

Image compression (IC) plays an important part in Digital Image Processing (DIP), it is as well very very essential for effective transmission and storing of images. Image Compression (IC), is basically recusing the size of an image and that too without adjusting the quality of the picture. It is kind of software with records pressure on digital Image. The objective is to lessen reiteration of the picture info for you to be accomplished of store or transmit information in a proficient shape. This paper gives review of kinds of images and its compression strategies. An image, in its genuine form, conveys big extent of data which requiress no longer finest large quantity of memory provisions for its garage but moreover causes difficult transmission over limited bandwidth channel. So, one of the acute factors for picture storage space or transmission over any exchange media is Image Compression. Image Compression makes it possible for increasing file sizes of practicable, storable and communicable dimensions.

IEEE Transactions on Information Theory, 2005

We present a new method for clustering based on compression The method doesn't use subject - specific features or background knowledge, and works as follows: First, we determine a universal similarity distance, the normal - ized compression distance or NCD, computed from the lengths of compressed data files (singly and in pairwise concatenation) Second, we apply a hierarchical clustering

iaeme

Use of Digital Image Communication has increased exponentially in the recent years. Joint Photographic Experts Group (JPEG) is the most successful still image compression standard for bandwidth conservation. Evidently JPEG Compression system consists of a DCT transformation unit followed by a quantizer and encoder unit. At the decoder end, image is created by inverse DCT. In this paper, we present a set of new JPEG Compression algorithms that combines K-Means clustering algorithm and DCT to further reduce the bandwidth requirements. Experiments are carried out with many standard still images. Our algorithm identified to be giving almost same Peak Signal-to-Noise Ratio (PSNR) as the standard JPEG algorithm.