This study proposes an effective pseudo relevance feedback method for information retrieval in the context of question answering. The method separates two retrieval models to improve the precision of initial search and the recall of feedback search. The topic-preserving query expansion links the two models to prevent the topic shift.

To study the similarity between queries and retrieved masses, we design an interactive CBIR (Content-based Image Retrieval) CADx (Computer-aided Diagnosis) system using relevance feedback for the characterization of breast masses in ultrasound (US) images based on radiologists' visual similarity assessment. The CADx system retrieves masses that are similar to query masses from a reference library based on six computer-extracted features that describe the texture, width-to-height, and posterior shadowing of the mass. The k-NN retrieval with Euclidean distance similarity measure and the Rocchio relevance feedback algorithm (RRF) are used. To train the RRF parameters, the similarities of 1891 image pairs from 62 (31 malignant and 31 benign) masses are rated by 3 MQSA (Mammography Quality Standards Act) radiologists using a 9-point scale (9=most similar). The best RRF parameters are chosen based on 3 observer experiments. For testing, 100 independent query masses (49 malignant and 51 benign) and 121 reference masses on 230 (79 malignant and 151 benign) images were collected. Three radiologists rated the similarity between the query masses and the computer-retrieved masses. Average similarity ratings without and with RRF were 5.39 and 5.64 for the training set and 5.78 and 6.02 for the test set, respectively. Average AUC values without and with RRF were, respectively, 0.86±0.03 and 0.87±0.03 for the training set and 0.91±0.03 and 0.90±0.03 for the test set. On average, masses retrieved using the CBIR system were moderately similar to the query masses based on radiologists' similarity assessments. RRF improved the similarity of the retrieved masses.

Kernel biased discriminant analysis (KBDA), as a subspace learning algorithm, has been an attractive approach for the relevance feedback in content-based image retrieval. Its performance, however, still suffers from the “small sample learning” problem and “kernel learning” problem. Aiming to solve these problems, in this paper, we present a new semi-supervised scheme of KBDA (S-KBDA), in which the projection learning and the “kernel learning” are interweaved into a constrained optimization framework. Specifically, S-KBDA learns a subspace that preserves both the biased discriminant structure among the labeled samples, and the geometric structure among all training samples. In kernel optimization, we directly optimize the kernel matrix, rather than a kernel function, which makes the kernel learning more flexible and appropriate for the retrieval task. To solve the constrained optimization problem, a fast algorithm based on gradient ascent is developed. The image retrieval experiments are given to show the effectiveness of the S-KBDA scheme in comparison with the original KBDA, and the other two state-of-the-art algorithms.

Localized content-based image retrieval (LCBIR) has emerged as a hot topic more recently because in the scenario of CBIR, the user is interested in a portion of the image and the rest of the image is irrelevant. In this paper, we propose a novel region-level relevance feedback method to solve the LCBIR problem. Firstly, the visual attention model is employed to measure the regional saliency of each image in the feedback image set provided by the user. Secondly, the regions in the image set are constructed to form an affinity matrix and a novel propagation energy function is defined which takes both low-level visual features and regional significance into consideration. After the iteration, regions in the positive images with high confident scores are selected as the candidate query set to conduct the next-round retrieval task until the retrieval results are satisfactory. Experimental results conducted on the SIVAL dataset demonstrate the effectiveness of the proposed approach.

In previous studies, the retrieval accuracy of large image databases has been improved as a result of reducing the semantic gap by combining the input sketch with relevance feedback. A further improvement of retrieval accuracy is expected by combining each stroke, and its order, of the input sketch with the relevance feedback. However, this leaves as a problem the fact that the effect of the relevance feedback substantially depends on the stroke order in the input sketch. Although it is theoretically possible to consider all the possible stroke orders, that would cause a realistic problem of creating an enormous amount of data. Consequently, the technique introduced in this paper intends to improve retrieval efficiency by effectively using the relevance feedback by means of conducting data mining of the sketch considering the similarity in the order of strokes. To ascertain the effectiveness of this technique, a retrieval experiment was conducted using 20,000 images of a collection, the Corel Photo Gallery, and the experiment was able to confirm an improvement in the retrieval efficiency.

Using query-by-sketch we propose an application to efficiently create collages with some user interaction. Using rough color strokes that represent the target collage, images are automatically retrieved and segmented to create a seamless collage. The database is indexed using simple geometrical and color features for each region, and histograms that represent these features for each image. The image collection is then queried by means of a simple paint tool. The individual segments retrieved are added to the collage using Poisson image editing or alpha matting. The user is able to modify the default segmentations interactively, as well as the position, scale, and blending options for each object. The resulting collage can then be used as an input query to find other relevant images from the database.

Content-based image search has long been considered a difficult task. Making correct conjectures on the user intention (perception) based on the query images is a critical step in the content-based search. One key concept in this paper is how we find the user preferred low-level image characteristics from the multiple positive samples provided by the user. The second key concept is how we generate a set of consistent "pseudo images" when the user does not provide a sufficient number of samples. The notion of image feature stability is thus introduced. The third key concept is how we use negative images as pruning criterion. In realizing the preceding concepts, an image search scheme is developed using the weighted low-level image features. At the end, quantitative simulation results are used to show the effectiveness of these concepts.

In this paper, we present our investigation of Latent Semantic Indexing (LSI) on the local query regions for solving the computation restrictions of the LSI on the global information space. Through the experiments with different SVD dimensionality on the local query regions, the results show that low-dimensional LSI can achieve much better precision than VSM and similar precision to global LSI. Such small SVD factors indicate that there is an almost linear surface in the local query regions. The largest or the two largest singular vectors have the ability to capture such a linear surface and benefit the particular query. In spite of the fact that Local LSI analysis needs to perform the Singular Value Decomposition (SVD) computation for each query, the surprisingly small requirements of the SVD dimension resolve the computation restrictions of LSI for large scale IR tasks. Moreover, on the condition that several relevant sample documents are available, application of low dimensional LSI for these documents can obtain comparable precision with the Local RF in a different manner.

This paper discusses a new type of semi-supervised document clustering that uses partial supervision to partition a large set of documents. Most clustering methods organizes documents into groups based only on similarity measures. In this paper, we attempt to isolate more semantically coherent clusters by employing the domain-specific knowledge provided by a document analyst. By using external human knowledge to guide the clustering mechanism with some flexibility when creating the clusters, clustering efficiency can be considerably enhanced. Experimental results show that the use of only a little external knowledge can considerably enhance the quality of clustering results that satisfy users' constraint.

In content-based image retrieval (CBIR), the content of an image can be expressed in terms of different features such as color, texture, shape, or text annotations. Retrieval methods based on these features can be varied depending on how the feature values are combined. Many of the existing approaches assume linear relationships between different features, and also require users to assign weights to features for themselves. Other nonlinear approaches have mostly concentrated on indexing technique. While the linearly combining approach establishes the basis of CBIR, the usefulness of such systems is limited due to the lack of the capability to represent high-level concepts using low-level features and human perception subjectivity. In this paper, we introduce a Neural Network-based Image Retrieval (NNIR) system, a human-computer interaction approach to CBIR using the Radial Basis Function (RBF) network. The proposed approach allows the user to select an initial query image and incrementally search target images via relevance feedback. The experimental results show that the proposed approach has the superior retrieval performance over the existing linearly combining approach, the rank-based method, and the BackPropagation-based method.