Content Based Image Retrieval (CBIR) is the process of retrieving visually similar images from huge datasets. Images are identified based on their content. Content identification using shape features is considered in this paper. Content identification using shapes is a challenging task considering multiple variations observed in images, complex backgrounds and vast categories of contents. This paper describes a shape descriptor based CBIR system. The content of an image is identified using a key point based shape descriptor. Template matching techniques are adopted to accurately describe object shapes. The object shape identified is described using histogram vectors. The use of SVM classifier for content recognition and image retrieval task is considered. Results presented prove robustness of the key point technique to accurately describe object shapes even in complex images. Performance of the proposed system is compared with existing state of art systems. Results obtained and described in the paper prove a better performance of proposed CBIR system.

Image based content recognition and retrieval is critical in many applications. Existing mechanisms for content based image retrieval lack in terms of performance. In this paper a hierarchical template tree based CBIR system is described. Content in image is represented using a combination of shape features and low level features. Comprehensive feature set definitions proposed enables in achieving better performance. Shape and low level features are considered as templates. Templates of similar categories are further decomposed to form a hierarchical template tree. Query image is converted into a query template and is decomposed. A part template based matching scheme and SVM classifier is used to retrieve visually similar images. Results presented in the paper prove superior performance of proposed technique when compared to recent existing mechanisms in place. An improvement of 10.45% and 9.69% in mean average precision and mean retrieval accuracy is reported using proposed approach.

No abstract

Glaucoma is a habitual eye disorder which harms eye’s second cranial nerve. There are millions of second cranial nerves. The main function of these types of nerves is to sending captured visual information from retina to the brain. The escalate pressure in the human eye leads to Glaucoma. This heavy pressure is known as intraocular pressure. This heavy pressure leads to damage eye's optic nerve head and retina continuously further it tends to vision loss. In this paper there are two datasets including both normal person and affected person's eye color images. The principal aim of this project is to compare the color of the eye with these two datasets. A special camera which is attached to less power microscope is called fundus camera or retinal camera. The images captured by this type of fundus camera is called fundus picture[1]. It is a high dimensional laser image. MATLAB software tool is used to fulfill the feature extraction of these fundus images. A color pixel in the affected area of the person is measured to check whether a person is Glaucomatous or not. If the final result is positive then it is Glaucoma.

In India agriculture is the main source of income for generating the economy. Diseases in plants are a major unavoidable problem, and hence detecting the diseases is the necessity of the day in the domain of agriculture. The main diseases found in tomato plants are viral, fungus and bacterial diseases. The detection will help improve the quantity and quality of the products with an optimum yield. In this paper a comparative analysis is carried out for the algorithms Support Vector Machine, Convolution Neural Networks, Decision tree classifier, and k-Nearest Neighbour (k-NN) with the result of 97%,97%,90% and 80% respectively.