Interactive Image Retrieval System Research Articles

Towards brain-activity-controlled information retrieval: Decoding image relevance from MEG signals

We hypothesize that brain activity can be used to control future information retrieval systems. To this end, we conducted a feasibility study on predicting the relevance of visual objects from brain activity. We analyze both magnetoencephalographic (MEG) and gaze signals from nine subjects who were viewing image collages, a subset of which was relevant to a predetermined task. We report three findings: i) the relevance of an image a subject looks at can be decoded from MEG signals with performance significantly better than chance, ii) fusion of gaze-based and MEG-based classifiers significantly improves the prediction performance compared to using either signal alone, and iii) non-linear classification of the MEG signals using Gaussian process classifiers outperforms linear classification. These findings break new ground for building brain-activity-based interactive image retrieval systems, as well as for systems utilizing feedback both from brain activity and eye movements.

Evaluation of Light Inspired Optimization Algorithm-Based Image Retrieval

Image retrieval system becoming a more popular in all the disciplines of image search. In real-time, interactive image retrieval system has become more accurate, fast and scalable to large collection of image databases. This paper presents a unique method for an image retrieval system based on firefly algorithm, which improve the accuracy and computation time of the image retrieval system. The firefly algorithm is utilized to optimize the image retrieval process via search for nearly optimal combinations between the corresponding features as well as finding out approximate optimized weights for similarities with respect to the features. The proposed method is able to dynamically reflect the user’s intention in the retrieval process by optimizing the objective function. The Efficiency of the proposed method is compared with other existing image retrieval methods through precision and recall. The performance of the method is experimented on the Corel and Caltech database images.

Class-specific metrics for multidimensional data projection applied to CBIR

Content-based image retrieval is still a challenging issue due to the inherent complexity of images and choice of the most discriminant descriptors. Recent developments in the field have introduced multidimensional projections to burst accuracy in the retrieval process, but many issues such as introduction of pattern recognition tasks and deeper user intervention to assist the process of choosing the most discriminant features still remain unaddressed. In this paper, we present a novel framework to CBIR that combines pattern recognition tasks, class-specific metrics, and multidimensional projection to devise an effective and interactive image retrieval system. User interaction plays an essential role in the computation of the final multidimensional projection from which image retrieval will be attained. Results have shown that the proposed approach outperforms existing methods, turning out to be a very attractive alternative for managing image data sets.

適合画像の自動分類による多グループフィードバックを用いた対話型画像検索

We previously proposed an interactive image retrieval system that uses the group-based relevance feedback method. We now propose a new system that groups relevant images automatically to reduce the burden on the user that existed in the previous system. We also previously introduced an efficient learning method that uses the rough set theory's “local reduct”. The results of our experiments demonstrated that, in our proposed system, the burden on the user is reduced without any decrease in retrieval performance. The proposed system's retrieval response speed and the suitable interface for trial-and-error retrieval are easy to use.

Image Information Mining System Evaluation Using Information-Theoretic Measures

During the last decade, the exponential increase of multimedia and remote sensing image archives, the fast expansion of the world wide web, and the high diversity of users have yielded concepts and systems for successful content-based image retrieval and image information mining. Image data information systems require both database and visual capabilities, but there is a gap between these systems. Database systems usually do not deal with multidimensional pictorial structures and vision systems do not provide database query functions. In terms of these points, the evaluation of content-based image retrieval systems became a focus of research interest. One can find several system evaluation approaches in literature, however, only few of them go beyond precision-recall graphs and do not allow a detailed evaluation of an interactive image retrieval system. Apart from the existing evaluation methodologies, we aim at the overall validation of our knowledge-driven content-based image information mining system. In this paper, an evaluation approach is demonstrated that is based on information-theoretic quantities to determine the information flow between system levels of different semantic abstraction and to analyze human-computer interactions.

ラフ集合論の縮約による多グループフィードバックを用いた対話型画像検索

In this paper, we propose an interactive image retrieval system that uses a novel relevance feedback method called group-based relevance feedback. In the proposed system, the user divides the relevant images into multiple groups according to his/her perspective. Using each user's perspective, the retrieval intention of the user will be captured more accurately than is possible by conventional methods, which use only images for retrieval. Moreover, the retrieval results are shown as grouped images, which facilitates the understanding of the user as to why such results were produced by the system. In order to implement the proposed system, we also introduce an efficient learning method that uses the “Reduct” from the Rough Set Theory to learn the retrieval intention of the user. Finally, retrieval results are presented in order to demonstrate the effectiveness of the proposed system and learning method.

An interactive image retrieval system with self‐explanation function

AbstractIn an image retrieval system, accurately extracting the content of the image which users want to search for based on an image in their heads is vital. As a result, many attempts have been made to use an image similar to the sought‐for image as a retrieval key and thereby to increase the freedom of expression for users' images. Because of the high variability of image characteristics, however, the vagueness of the content which indicates the retrieval key itself rises compared to conventional methods that use keywords or other linguistic resources for the retrieval key. Therefore, in this research the authors consider these problems by reviewing image retrieval from the perspective of image communications between users and the system, and then propose an interactive mechanism which limits the increase in vagueness while using an image as the retrieval key. The authors then perform retrievals by having users revise the retrieval key based on this mechanism. The authors create their interactive mechanism in a real image retrieval system and confirm its validity through experiments. © 2001 Scripta Technica, Syst Comp Jpn, 32(13): 41–51, 2001