Close Clustering Based Automated Color Image Annotation - Computer Science > LearningReportar como inadecuado




Close Clustering Based Automated Color Image Annotation - Computer Science > Learning - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Abstract: Most image-search approaches today are based on the text based tagsassociated with the images which are mostly human generated and are subject tovarious kinds of errors. The results of a query to the image database thus canoften be misleading and may not satisfy the requirements of the user. In thiswork we propose our approach to automate this tagging process of images, whereimage results generated can be fine filtered based on a probabilistic taggingmechanism. We implement a tool which helps to automate the tagging process bymaintaining a training database, wherein the system is trained to identifycertain set of input images, the results generated from which are used tocreate a probabilistic tagging mechanism. Given a certain set of segments in animage it calculates the probability of presence of particular keywords. Thisprobability table is further used to generate the candidate tags for inputimages.



Autor: Ankit Garg, Rahul Dwivedi, Krishna Asawa

Fuente: https://arxiv.org/







Documentos relacionados