Region Based Image Retrieval Approach Using SVM with Shape and Texture Based Feature

Abstract

The content based image retrieval is important research in medical and internet applications. This aim is searching the images from database based on query image. Region based image retrieval approach using edge following technique with shape and texture based feature are present in this paper. The Gradient based cluster algorithm is used to find the boundaries of noisy medical images. The performance and robustness of edge following technique is tested to detect the objects in different medical images. To accomplish this, local binary pattern (LBP) as texture feature and Hu moments (Hu), Tchybysheve Moments, invariant moment as shape based features are taken in this paper. These shape and texture features are calculated from segmented images and combine as feature vectors. The advantage of feature vector is to reduce the image dimensionality and less computation time. After feature vector is calculated, comparing the feature vector of query image and feature vectors of given database images and then sorted the images based on distance. The proposed method has better to improve the retrievals accuracy, high speed and less computation time. Our obtained results show the proposed method is robust in different medical image alterations. The proposed method obtains 99% retrieval accuracy, which is satisfactory. The region based image retrieval performance of proposed method is tested of 4000 images with 40 different classes.

Authors and Affiliations

Bharath. P, Prasad. C

Keywords

Related Articles

slugSurvey of Artificial Electronic Nose to Detect the Diseases of Human

The main purpose of this paper is to identify the diseases of people and give the health status of people in less time and in low cost. On using artificial electronic nose. Artificial electronic nose is an electronic de...

Design and Optimization of Multiple Slots Microstrip Patch Antenna for Wireless Applications

Multiband Wang-shaped fractal antenna is obtained by applying fractal geometry .Initially a rectangular patch is taken and fractal geometry is applied. Two iterations of fractal geometry are applied to form double wang-...

slugCloud Computing: A Beginners Primer

Information Technology (IT) has embarked on a new paradigm — cloud computing. Cloud computing provides alternative services to deliver computer resources and has sparked a revolution in the way organizations provide inf...

Substation Automation System for Energy Monitoring and Control using SCADA

Inherent scarcity of thermal sources with a time varying temperature profile is the reason why the Thermo electricbased energy harvesting is not as prominent as its counter-part, thermoelectric generators, in the thermal...

Performance and Emission Characteristics of S.I. Engine with Oxygenates

In modern world transportations play a vital role, in that S.I. engines are the mostly preferred for their luxurious transportation. Alternative fuels are prepared to trim down the dependency of gasoline as a fuel and i...

Download PDF file
  • EP ID EP22953
  • DOI -
  • Views 236
  • Downloads 5

How To Cite

Bharath. P, Prasad. C (2016). Region Based Image Retrieval Approach Using SVM with Shape and Texture Based Feature. International Journal for Research in Applied Science and Engineering Technology (IJRASET), 4(12), -. https://europub.co.uk/articles/-A-22953