REVIEW ON THINNING TO INCREASE THE
THINNING RATE OF BINARY IMAGE
Ramneet Kaur
1, Er. Amritpal kaur
2,
1
M.Tech Researcher Student,
2Assistant Professor,Department of CSE,
Sri Guru Granth Sahib World University(Fatehgarh Sahib)(India)
ABSTRACT
Thinning is the preprocessing stage to make easy higher level analysis and recognition for such applications
like OCR, thinning techniques have been applied in many fields such as automated industrial inspection, pattern
recognition, biological shape description and image coding etc. The main objective of thinning is to improve
efficiency, to reduce transmission time. There are various algorithms to improve thinning rate. It is concluded
that there are some loopholes in thinning algorithm. So there is a need to improve thinning rate .In the proposed
enchancement in Zhang-suen algorithm to improve thinning rate.
I. INTRODUCTION
Digital Image Processing comprises of three words: Digital, Image & Processing. Image processing is a method
to perform some operations on an image, in order to get an enhanced image or to extract some useful
information from it. By transforming the binary image into skeleton which reduce the storage of data and reduce
the transmission time also. Skeletonization is the process of extracting skeletons from an object in a digital
image. It is morphological operation that deletes black foreground pixels iteratively layer by
layer until one pixel width skeleton is obtained. There are many image analysis techniques in which
skeletonization is a “pre-processing” step [2]. It is a process of
reducing an object in a digital image to the minimum size necessary for machine recognition of that object
[3].Binary image is obtain by the combination of black and white pixels. By converting the input binary image
into output .
II. APPLICATIONS OF THINNING
Skeletonization has been used for variety of image processing applications like:
Fingerprint classification
Biometric authentication
Signature verification
Medical imaging
III. RELATED WORK
3.1 Homotopic thinning
The homotopic thinning is a morphological image trasformation that aims to represent object by their medical
axis lines while preserving image topology. Homotopic thinning is applied on binary image which proceed by
removing the pixels called simple pixels. By removing the border pixels while not change the background nor
the foreground connectedness.Simple pixels configuration is more important than introduce end pixel notation
to obtain the thin skeleton that preserve the extremities as well as the topology. Infact, adapting the binary
thinning to the grayscale case by preprocessing and thresholding may cause irrevocable information loss. The
gray skeleton is obtained after stability of an iterative peeling consisting here in lowering gray values. It is
obtain by the composed of pixels located at the crest lines. We mainly focused on fitting the thinning parameter
for removal of insignificant information and this approach implement for considering the connectedness
preservation .Firstly leads to the adjustment of thinning parameters and secondly leads to overcome the
algorithms challenges related to the ascendant gray level treatment by using hierarchical queues.
3.2 Grayscale thinning
Binary image consist of background pixels({x ,F(x)=0}) and foreground pixels ({x ,F(x)=1}) also known as
object pixels . Homotopic thinning is applied on binary image to remove the object pixels without changing the
connectedness of both foreground and background pixels .The 4connectivity N4 is classically used for
background and the 8 connectivity N8 for the foreground. There are two conditions and the first condition states
simply that x is a border object and second condition that removing x doesn’t change the object connectedness.
This process obtain thin skeleton composed only of barrier between connected components with no unblocked
branches. The end pixels are introduced in order to protect extremities was also extended to gray tone images.
The behavior of thinning algorithms for noisy images are leads to over connected skeletons. Then thinning
measurement does not allow disconnection and dissociate noise related and significant information.
3.3 Parameter based thinning
For the noisy image result show a lot of non significant crests .To improve this a local contrast parameter
introduce called The thinning constraints are relax by parametric thinning by lowering low contrast crests, peak
and ends according to manually selected parameters. Parametric thinning process having the local adjustment
and standardization .The thinning parameters by using the statistical adjustment linked to both noise standard
deviation and image contrast .Result of statistical adjustment and implementation illustrated on both synthetic
IV. REVIEW OF LITERATURE
Abu-Ain W, et al. in this paper a new skeletonization algorithm was proposed to combine both iterative parallel
and sequential approaches .This algorithm Is made up of three stages .Skeletons are obtain using first two steps
and third one is used to optimizing the extracted skeletons into single pixel width. An experimental result gives
better results than the previous skeletonization algorithms.
Padole G.V.Pokle S.B. describe two iterative algorithm for the binary image thinning. In the first given
algorithm, two operations are described edge detection and subtraction are used for the thinning of binary
image. Second algorithm is based on repeatedly deleting the pixels until a one pixel thick pattern in a binary
image is obtained. Erosion conditions are devised to assure preserving connectivity. Experimental results show
that edge based iterative thinning algorithm is time consuming as compared to optimized skeletonization
algorithm.
T. Zhang et al. in this paper thinning of binary images is done by repeating two sub-iterations in which one
delete the south-east boundary points and the west corner points while the second is for delete the
north-west boundary points and south-east corner points .The point are deleting according to a specific set of rules.
The two sub-iterations are repeated until no more points validate the deleting rules.
Shang L., Yi Z. a new sequential algorithm is proposed in which uses flag map and bitmap simultaneously to
decide whether a boundary pixel should be deleted or not. To compare the previous algorithm with other there
are three performance criteria are proposed in this paper. Experimental results given by the proposed algorithm
is not only one pixel thick and perfectly connected, well defined but are also immune to noise.
Datta A. et al. mainly two new parallel algorithms are proposed in this paper. Those algorithm convert the image
into one pixel thick width and help us to maintain the 8neighbour connectivity in binary image. This proposed
algorithm gives the better result than the previous algorithms.
Ahemed P. in this paper a new K3M skeletonization algorithm was proposed that show some interesting
properties related to processing quality and algorithm clarity, enriched with examples.
Li Z. et al. to maintain the connectivity of the binarized fingerprint image a new robust parallel thinning
algorithm is proposed to obtain the skeleton which is one pixel wide which gets extremely close to the medical
axis .Three sub iterations are repeated by proposed thinning method .Result gives by proposed robust parallel
thinning algorithm obtain better skeletons than the previous algorithm.
Kumar V. et to make the algorithm automatic the proposed algorithm measure the value of connected
component. In this algorithm there is no requirement of human interaction and also free from the shape and font
and does not required any preprocessing .So it is better than others.
The following table1 shows the advantage and disadvantages of different thinning techniques
Table 1. Comparison between different thinning techniques
Author Description Advantages Disadvantages
Padole G.V
(2010)
Two new iterative
algorithms are proposed
T. Zhang et al.
(1984 )
Perform thinning by
repeating two sub-iteration
Efficient, fast and flexible Problem in reliable
segmentation
Ahmed P.
(2002)
Propose new
skeletonization algorithm
K3M
Symmetrical thinning,
speed and thin symbol of
any language
Expansive
Datta A. et al. Shape extraction Grid independent Noise sensitivity and
rotation dependency
Li Z. et al. Propose a robust parallel
thinning algorithm
Remove unexpected edge
disconnection
Problem in the quality of
skeleton
Kumar V.et al. Describe new novel
scheme for thinning
Measure value of
connected component, it is
automatic
Overcome loss of
information problem
Abu-ain W, et al.
(2013)
Combine sequential and
parallel approaches
Reduction of memory
space required to store
essential structure
Some portion of same is
totally disapperared
V. CONCLUSION
This paper provides the survey of various algorithms for binary images thinning. The main aspects of
skeletonization algorithms is noise immunity. In the proposed work by the use of neural network and Zhang and
Suen algorithm we make the algorithm more robust than the previous skeletonization techniques. By comparing
the existing techniques with the proposed techniques in the terms of parameters such as execution time, memory
,thinning rate, number of connected components, PSNR,MSE etc which makes the proposed algorithm better
than the previous skeletonization algorithms.
REFERENCES
[1] Abu-Ain W, et al. “Skeletonization Algorithm for Binary Images” The 4th International Conference
onElectrical Engineering and Informatics (ICEEI 2013) pp 704-709
[2] Padole G.V, Pokle S. B. “New Iterative Algorithms For Thinning Binary Images” Third International
Conference on Emerging Trends in Engineering and Technolgy IEEE 2010 pp. 166-171
[3] Datta A. et al “Shape Extraction: A rative Study Between Neural Network-Comparative Study
Between Neural Network-Based and Conventional Techniques” Neural Computing & Applications (1998)
Springer pp. 343-355
[4] T. Zhang et al. “A fast parallel algorithm for thinning digital patterns”, Commun. ACM 27 (3) (1984) pp.
236-239
[5] Ahmed et al. “A Rotation Invariant Rule-Based Thinning Algorithm for Character Recognition” IEEE
[6] Kumar V. et al. “A New Skeletonization Method Based on Connected Component Approach” IJCSNS
International Journal of Computer Science and Network Security, VOL.8 No.2, February 2008 pp. 133-137
[7] Li Z. et al. “Modified Binary Image Thinning Using Template-Based PCNN” (2013) International
conference on information technology and software engineering volume 212 pp.731 -740
[8] Shang L., Yi Z. “ A class of binary images using two PCNNs” Neurocomputing 70(2007) pp.