Effective Graph-based Content-Based Image Retrieval Systems for Large-Scale And Small-Scale Image Databases

Distance measures like the Euclidean distance are used to measure similarity between images in content-based image retrieval. Such geometric measures implicitly assign more weighting to features with large ranges than those with small ranges. This paper discusses the effects of five feature normalization methods on retrieval performance. We also describe two likelihood ratio-based similarity me...

Content Based Image Retrival needs relevance feedback to obtain more accurate results. By building a repository with KBDA, very good results may be obtained with the use of a Query Semantic Feature Vector.

Over the last few years the interest in the research problem of indoor vs. outdoor scene classification[2] has grown significantly, due to its importance in many applications such as Content Based Image Retrieval (CBIR) or Query by Image Content (QBIC), image data organization, robotics, and photography, thus providing a strong motivation for this project. For an example, knowledge of a scene t...

Visual relationships capture a wide variety of interactions between pairs of objects in images (e.g. “man riding bicycle” and “man pushing bicycle”). Consequently, the set of possible relationships is extremely large and it is difficult to obtain sufficient training examples for all possible relationships. Because of this limitation, previous work on visual relationship detection has concentrat...

In content-based image retrieval (CBIR), the effectiveness of the low-level features depends on their capabilities in describing the high-level semantic concepts. How to properly evaluate such an effectiveness remains a challenge. In this paper, we address the evaluation problem by using the decisive feature patterns of the low-level features. Intuitively, a decisive feature pattern is a combin...

Texture is commonly used feature in most of the content-based image retrieval systems. This texture retrieval ability can be also applied to rock texture. The retrieval of the rock texture is a demanding task because of special character of rock. In this paper some existing contentbased image retrieval systems are tested with a sample set representing clearly different rock images. The recall a...

Image retrieval has great potential for a variety of tasks in medicine but is currently underdeveloped. For the ImageCLEF 2005 medical task, we developed a context-based retrieval system. We then conducted our experiment using an automatic query, a manual query, and a manual/visual query. The best results were obtained from manual modification of queries, while combining those results with data...

This paper presents a multi-mode indexing scheme for effective content-based image retrieval. Three types of indices are identified: visual indices for quantifiable visual information, semantic indices for non-quantifiable semantic information, keywords indices for keywords or free text. The underlying index structures are the HG-tree and the signature file. The HG-tree is one of the most promi...