Fast indexing: a comparative evaluation
نویسندگان
چکیده
In this evaluation the efficiency of three image signature called angular spectrum, Hough based signature and color histogram are tested The first signature is intrinsic hierarchical (deriving from image frequency spectrum) and than non signature space reduction technique is used. The second is a short signature directly indexed and the last (color histogram) need to be reduced for fast indexing. The best performance (efficiency and effectiveness) is obtained using angular spectrum. Key-Words: high-dimensional indexing and retrieval, similarity search, image databases, FFT, Color Histogram, Hough transform
منابع مشابه
Comparative Evaluation of Data Stream Indexing Models
to the user's queries should be time and space efficient. In this paper, we consider the special requirements of indexing to determine the performance of different techniques in data stream processing environments. Stream indexing has main differences with approaches in traditional databases. Also, we compare data stream indexing models analytically that can provide a suitable method for stream...
متن کاملNotes on the Limits of CLIR Effectiveness: NTCIR-2 Evaluation Experiments at Justsystem
NTCIR-2 evaluation experiments at the Justsystem site are described with a focus on comparative study of CLIR effectiveness with monolingual retrieval effectiveness of the same retrieval engine. Experiments on the effects of phrasal translation, indexing of translated phrasal terms, pre-translation feedback and parallel documents feedback in diverse retrieval settings, are reported. The results...
متن کاملمدل دو مرحله ای شکاف- گلچین برای نمایه سازی خودکار متون فارسی
Purpose: Each language has its own problems. This leads to consider appropriate models for automatic indexing of every language. These models should concern the exhaustificity and specificity of indexing. This paper aims at introduction and evaluation of a model which is suited for Persian automatic indexing. This model suggests to break the text into the particles of candidate terms and to c...
متن کاملA Comparative Study on the Use of Multi-label Classification Techniques for Concept-Based Video Indexing and Annotation
Exploiting concept correlations is a promising way for boosting the performance of concept detection systems, aiming at conceptbased video indexing or annotation. Stacking approaches, which can model the correlation information, appear to be the most commonly used techniques to this end. This paper performs a comparative study and proposes an improved way of employing stacked models, by using m...
متن کاملBulkloading of Bitmap Indices
The conventional indexing technique designed for the traditional database systems does not use the read-only characteristics of the data warehouse environment. Recently, Bitmap Indexing has been proposed as a candidate indexing structure suitable for the read-only environments, like those of decision support systems. The current paper presents a short literature survey of the various clone of t...
متن کامل