The resource usage in Content-Based Image Retrieval is a frequently neglected issue. This paper describes a novel compact feature vector based on image color histograms in the HSL color space. The images are represented using only 10 bytes per image. It is shown that, in the context of Query-by-Example (QbE) usage scenarios, the method described achieves retrieval performance close to the state of the art image retrieval methods that use considerably more memory. It is also shown that the described method outperforms other methods with similar memory usage.

Color histograms are one of the earliest and best known image features used in Content-Based Image Retrieval (CBIR). There is a wealth of scientific work on this topic. However, different papers vary in the specific ways of determining histograms and distance between them. In this paper authors attempt to classify various types of histograms used in literature and compare them using contemporary datasets and metrics for evaluation. Histograms are compared based on their retrieval performance as well as resource usage.

In this paper we describe a case based approach to group recommendation process in which more than one person is involved in the recommendation process. The main problem group recommendation needs to solve is how to adapt to the group as a whole based on item features describing individual user preferences. Our approach takes into account that the distribution of individually perceived feature importance weights variate among members of the group. The two methods to case representation are presented: case representation by combining individually perceived feature importance weights and case representation by combining averaged perceived feature importance weights. In order to compare these two methods to case representation, the two metrics widely used in information retrieval (recall and precision) are used.

Content-based image retrieval (CBIR) is a field of active research for almost 20 years. This timeframe has seen several generations of hardware and corresponding changes in computer usage patterns. It is therefore prudent to periodically reevaluate known methods in the context of modern hardware and usage patterns. Overall the issue of resource usage in CBIR is somewhat neglected. In this paper some extremes in this area are benchmarked and results presented. Specifically, paper is focused on usage scenario of indexing a personal image collection.

In this paper we describe a case based approach to product recommendation process in which contextual information are integrated. The integration of contextual information into recommender systems adds an additional information value to the recommender data model. In order to evaluate the effects of the integration of contextual information, the case representation with integrated contextual information is compared with the case representation without contextual information. We compare these two approaches to case representation in terms of the two metrics widely used in information retrieval (recall and precision).

The discrete cosine transform (DCT) is a widely used technique in content image retrieval. This transform concentrate signal energy into lower order coefficients. In this paper, the construction of DC feature vector and three type of simple binarized feature vectors generated from AC DCT coefficients is presented. Next, in order to capture spatial layout information, annular histograms of introduced feature vectors are described. Despite of the simplicity of introduced binarized feature vectors, experimental results shows that these vectors can contribute to the overall retrieval performances.

Iterative and incremental development (IID) is a staple of multiple software development methods, including Spiral development, Unified process and Agile development. However, a number of papers warn of possible pitfalls in application of this approach in examples of real-life development project. In this paper we will describe an (in our opinion) successful application of IID to a software project which is in everyday use and which has so far proven useful. In addition, we will attempt to draw some conclusions that may be of use for further research on software development methodology.

element method is a method for solving partial differential equations. The method is applicable to a wide range of physical and engineering problems that can be mathematically described by partial differential equations. The STFFEM's learning material organization has a hierarchical structure with four levels. The STFFEM is able to utilize the specific knowledge of previously experienced, concrete content sequencing situations (cases). A new content sequencing problem (a new case) is solved by finding a similar past case, and reusing it in the new content sequence problem situation.

previous interaction experience represented in the form of cases. Cases include observations of the virtual environment as well as behavior routines that represent the solution component of cases. The AVA that continuously interacts with a virtual environment must be able to autonomously create new situation cases based on its perception of the local virtual environment in order to select the appropriate actions to achieve the goal state.

This paper presents an approach to agent's action selection strategy called stepwise case-based reasoning (SCBR). In this paper an agent that roams in 3D synthetic world is called an SCBR agent. An SCBR agent is an entity that selects the next action based on previous interaction experience and on simulated vision. An SCBR agent's interaction experience is represented in the form of two different types of cases: plan cases and contextual cases.