Integrated Computer-Aided Engineering

Level sets appear a suitable basis to decomposing an image into a meaningful organization of basic entities. Handling images by a basis of associated sets proves more effective than considering images as the collection of acquired intensity values, even in several frequency bands, due to the various noise sources. Then, level lines and sets are more stable features to support recognition than edges likely resulting from high pass filtering. Color being assumed more informative than intensity is preferred to exhibit the set basis. In the paper, color lines are defined after level lines and they are extracted and characterized consecutively. Greater information is conveyed by color lines resulting into more efficient grouping towards objects. A novel Hausdorff-inspired disparity finder is introduced fed in by color lines with respect to epipolar constraints. The efficient disparity map obtained by pixel wise line-matching between left and right images justifies our technical choices.

There exists a high interest in the digitization of handwriting historical documents, in the quest to preserve the cultural heritage of nations. In general, these manuscript images present new segmentation difficulties with respect to non-historical documents. The problems come from features such as paper aging, faded ink, back-to-front ink superposition or variable line skew, among others. This paper presents a methodology for detecting and extracting the text lines of images from complex handwritten historical documents. The proposed line segmentation algorithm is based on computing a binary transition map of the document and then extracting and refining the corresponding line regions through skeletonization. To improve the accuracy of line segmentation, a new graph-based splitting method to separate the touching lines is introduced. Once text lines have been segmented, we propose an algorithm based on mathematical morphology operators and position heuristics, to extract the component words on each text line. The robustness and accuracy of our approach was tested on digitalized pages of two complex historical document datasets: the correspondence of Nabuco and the family papers of Graham Bell. We have also successfully compared our algorithms to other general line and word segmentation algorithms presented at the ICDAR 2007 Handwriting Segmentation Contest.

Dynamic magnetic resonance imaging is an emerging technique for studying speech production. Vocal tract image sequences acquired during the speech of words or phonemes allow the identification of shapes taken by the vocal tract during speech production. However, there is no prior knowledge about the spatial and temporal resolution requirements, which are expected to vary depending on the speech task. Available approaches try to enhance resolution by empowering the acquisition devices. However, there are several cost and hardware limitations. In this paper, we propose an alternative approach to enhance spatio-temporal resolution using only digital image processing techniques. We use a previous non-rigid image registration method to identify displacements and deformations among the observed images. Based on a motion compensated interpolation approach, temporal resolution is increased by generating intermediate images which are coherent with the movement present in the observed sequence. Moreover, a MAP-MRF super-resolution image reconstruction approach is used to increase spatial resolution of the whole sequence. Results indicate the effectiveness of our approach.

Human beings are constantly stimulated to activate mental operations associated with emotion. Moreover, there are different neural mechanisms that correspond to each emotion degrees. Science lacks knowledge on the operation of such mechanisms, despite being notorious that some regions are related to particular emotions. Modern studies generally enumerate the relationship <mental process, activation region>, while it would be much more interesting if they explain the brain engine aspects. This article intends to present a technique for modeling human brain as a computable object by using Computer Theory and Digital Image Processing. A pattern cerebral activation recognizer automaton was constructed from functional magnetic resonance images in order to systemize the cerebral activation human process for a limited set of emotional stimulations.

There has been a growing interest in implementing image search engine at the semantic level. However, most existing practical systems including popular commercial image search engines like Google and Yahoo! are either text-based or a simple hybrid of texts and visual features. This paper proposes a novel image search system based on automatic image annotation. We develop a technology which learns semantic image concepts from image contents and transforms unstructured images into textual documents, so that images are indexed and retrieved in the same way as textual documents. Existing database management systems can be used to effectively manage image contents, and image search can be as efficient as text search by transforming images into textual documents through machine learning. Experiments in both the Corel dataset and real Web dataset are performed to validate our system and the results are promising. This system suggests a new combination of texts and visual features in order to achieve a semantic image search, and is expected to become a re-ranking system to the existing image search result via the Internet.

In recent years, how to extract useful information and knowledge from fuzzy relational databases has received much attention. Based on the high expressive power and effective reasoning service of Description Logics (DLs), this paper proposes a DL approach for automatically extracting knowledge from fuzzy relational databases (FRDB). To represent the extracted knowledge, a fuzzy DL called f-ALCNI is introduced after considering the characteristics of FRDB. On this basis, we propose an approach which can extract the f-ALCNI knowledge base from the FRDB, i.e., which can transform the FRDB (including schema and data information) into the f-ALCNI knowledge base (i.e., TBox and ABox). Furthermore, we design and implement a prototype extraction tool called FRDB2DL. In addition, to further demonstrate how the DLs are useful for improving some database applications, based on the extracted knowledge, we investigate the reasoning problems of FRDB (e.g., consistency, satisfiability, subsumption, equivalence, and redundancy) by means of the reasoning mechanism of f-ALCNI. Case studies show that the proposed approach is feasible and the tool is efficient.