Copyright © 2018 Authors. This is an open access article distributed under the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
International Journal of Engineering & Technology, 7 (2.22) (2018) 9-14
International Journal of Engineering & Technology
Website: www.sciencepubco.com/index.php/IJET
Research paper
Effective lossy and lossless color image compression with
Multilayer Perceptron
Dr. PL. Chithra *, A. Christoper Tamilmathi
Department of Computer science, University of Madras, Chennai, India
*Email: chitra.cs@unom.ac.in
Abstract
This paper presents the effective lossy and lossless color image compression algorithm with Multilayer perceptron. The parallel struc-
ture of neural network and the concept of image compression combined to yield a better reconstructed image with constant bit rate and
less computation complexity. Original color image component has been divided into 8x8 blocks. The discrete cosine transform (DCT)
applied on each block for lossy compression or discrete wavelet transform (DWT) applied for lossless image compression. The output
coefficient values have been normalized by using mod function. These normalized vectors have been passed to Multilayer Perceptron
(MLP). This proposed method implements the Back propagation neural network (BPNN) which is suitable for compression process
with less convergence time. Performance of the proposed compression work is evaluated based on three ways. First one compared the
performance of lossy and lossless compression with BPNN. Second one, evaluated based on different sized hidden layers and proved
that increased neurons in hidden layer has been preserved the brightness of an image. Third, the evaluation based on three different
types of activation function and the result shows that each function has its own merit. Proposed algorithm has been competed with
existing JPEG color compression algorithm based on PSNR measurement. Resultant value denotes that the proposed method well
performed to produce the better reconstructed image with PSNR value approximately increased by 21.62%.
Keywords: Activation function; Back propagation neural network; Discrete cosine transform; Error; Hidden layer; JPEG compression.
1. Introduction
Nowadays communication medium rapidly transformed with image
transformation. So the image compression has taken the essential
part to reduce not only the storage and band width of the transmis-
sion and also the time and cost. Image compression has been done
in two ways, lossy and lossless compression. JPEG compression
method is the one of the most significant lossy compression algo-
rithm. Another one is DWT for lossless image compression. This
proposed work deals with both JPEG and DWT compression com-
bined with multilayer perceptron. BPNN is one of the Multilayer
perceptron algorithms. Combination of BPNN and JPEG compres-
sion algorithm eliminates the need of quantization table and entropy
symbol table and produces the constant bit rate compression. The
Training algorithm and back propagation neural network is used to
increase the performance and to decrease the convergence time and
provide high compression ratio with low distortion [1]. Effectively
compress a wide range of novel images using back propagation.
The networks operate and are trained on residual image blocks [2].
Haar wavelet transform and discrete cosine transform are consid-
ered and a neural network is trained to relate the x-ray image which
controls compression method and their optimum compression ratio
[3]. Back propagation neural network algorithm helps to increase
the performance of the system and to decrease the convergence time
for the training of the neural network [4]. Fast BP results such as
compression ratio (CR) and peak signal to noise ratio (PSNR) are
computed and compared with BP results. From the results, we no-
ticed that the use of FBP improve the BPNN training by reducing
the convergence time of image compression learning process [5].
Neural network is well suited for real time systems because of their
fast response and computational times which are due to their paral-
lel architecture [6]. If a network gets trained up successfully with a
particular activation function, then there is a high probability that
other activation function will also lead to proper training of neural
network [7].
The concept of image compression using ANN and DWT possess
the advantages of simple computations, fault tolerance, parallel pro-
cessing, robust with respect to error transmission in the communi-
cation media which has made a breakthrough in supervised learning
of layered neural network [8]. Keeping the quality of the image con-
siderably same even after transmitting via noisy channel in their
BPNN algorithm. Thus the data transmission can be done with less
band width, power and storage space [9]. JPEG image compression
using FPGA with artificial neural networks gives more compres-
sion ratio when compared to existing systems [10]. The implement
of ANN with CSD based multiplier is thus proven to be one of the
effective processes for image compression and decompression [11].
Steps of back propagation algorithms and manual calculations have
been displayed [12]. In image compression technique convergence
time also play main role for quality of image and the back propaga-
tion neural network and Levengberg - marqurdt algorithms estimat-
ing a connection counted in which image compression and conver-
gence time have been improved [13]. The bipolar technique is pro-
posed and implemented for image compression and obtained the
better results as compared to principal component analysis (PCA
technique). It is observed that the Bipolar & LM algorithm suits the