Data Mining - Concepts and Techniques.

An adaptive interface for a programmable system, for predicting a desired user function, based on user history, as well as machine internal status and context. The apparatus receives an input from the user and other data. A predicted input is presented for confirmation by the user, and the predictive mechanism is updated based on this feedback. Also provided is a pattern recognition system for a multimedia device, wherein a user input is matched to a video stream on a conceptual basis, allowing inexact programming of a multimedia device. The system analyzes a data stream for correspondence with a data pattern for processing and storage. The data stream is subjected to adaptive pattern recognition to extract features of interest to provide a highly compressed representation which may be efficiently processed to determine correspondence. Applications of the interface and system include a VCR, medical device, vehicle control system, audio device, environmental control system, securities trading terminal, and smart house. The system optionally includes an actuator for effecting the environment of operation, allowing closed-loop feedback operation and automated learning.

Ergonomic man-machine interface incorporating adaptive pattern recognition based control system

Computer and Robot Vision Vol. 1, by R.M. Haralick and Linda G. Shapiro, Addison-Wesley, 1992, ISBN 0-201-10887-1.

Computer and Robot Vision

Recent advances in spectral-spatial classification of hyperspectral images are presented in this paper. Several techniques are investigated for combining both spatial and spectral information. Spatial information is extracted at the object (set of pixels) level rather than at the conventional pixel level. Mathematical morphology is first used to derive the morphological profile of the image, which includes characteristics about the size, orientation, and contrast of the spatial structures present in the image. Then, the morphological neighborhood is defined and used to derive additional features for classification. Classification is performed with support vector machines (SVMs) using the available spectral information and the extracted spatial information. Spatial postprocessing is next investigated to build more homogeneous and spatially consistent thematic maps. To that end, three presegmentation techniques are applied to define regions that are used to regularize the preliminary pixel-wise thematic map. Finally, a multiple-classifier (MC) system is defined to produce relevant markers that are exploited to segment the hyperspectral image with the minimum spanning forest algorithm. Experimental results conducted on three real hyperspectral images with different spatial and spectral resolutions and corresponding to various contexts are presented. They highlight the importance of spectral-spatial strategies for the accurate classification of hyperspectral images and validate the proposed methods.

/pdf/advances-in-spectral-spatial-classification-of-hyperspectral-vlt2ik4ef1.pdf

Advances in Spectral-Spatial Classification of Hyperspectral Images

An adaptive interface for a programmable system, for predicting a desired user function, based on user history, as well as machine internal status and context. The apparatus receives an input from the user and other data. A predicted input is presented for confirmation by the user, and the predictive mechanism is updated based on this feedback. Also provided is a pattern recognition system for a multimedia device, wherein a user input is matched to a video stream on a conceptual basis, allowing inexact programming of a multimedia device. The system analyzes a data stream for correspondence with a data pattern for processing and storage. The data stream is subjected to adaptive pattern recognition to extract features of interest to provide a highly compressed representation that may be efficiently processed to determine correspondence. Applications of the interface and system include a video cassette recorder (VCR), medical device, vehicle control system, audio device, environmental control system, securities trading terminal, and smart house. The system optionally includes an actuator for effecting the environment of operation, allowing closed-loop feedback operation and automated learning.

Adaptive pattern recognition based control system and method

Meter to sub-meter resolution satellite images have generated new interests in extracting man-made structures in the urban area. However, classification accuracies for such purposes are far from satisfactory. Spectral characteristics of urban land cover classes are so similar that they cannot be separated using only spectral information. As a result, there is an increased interest in incorporating geometrical information. In current literature, this is achieved by using an object-based approach. This requires a segmentation process. However, the complex objects in urban remote sensing images make this process very difficult. In this paper, we propose a method to measure the minimum and maximum dimension of an object, without however performing a segmentation. This method is based on morphological profiles (MP). Previous work on MP’s have shown the potential for improving classification results. However, an MP contains many values for each pixel, which can lead to problems of dimensionality. Feature extraction algorithms could reduce the dimensionality, but the resulting features are no longer interpretable. In this paper we use MP’s to derive a measure of minimum and maximum object dimension. These two measures allow to differentiate between long (roads) and more compact objects (buildings). We show that these new features improve the classification substantially.

Improved classification of urban areas with pixel-based measures of object size

Comparison between transport mode classifiers is usually performed without considering imbalanced samples in the dataset. This problem makes performance rates, such as accuracy and precision, not enough to report the performance of a classifier because they represent a cut-off point in the classifier performance curve. Our rule-based method proposes to combine both, the network elements associated with the transport mode to identify, and the elements associated with other means of transport. We performed a comparison between our proposed method and another GPS/GIS-based method, by applying a real-world representative dataset with a target class imbalance. We evaluated the performance of both methods with five experiments, using the area under the Receiver Operating Characteristic curve as metric. The results show that the tested methods achieve the same false positive rate. However, our method identifies correctly 84% of the true positive samples, i.e., the highest performance in our test data (data collected in Belgium). The proposed method can be used as a part of the post-processing chain in transport data to perform transport and traffic analytics in smart cities.

/pdf/efficient-use-of-geographical-information-systems-for-4gdx8fvrwd.pdf

Efficient use of geographical information systems for improving transport mode classification

Describing image content for quality assessment of geospatial data

This paper investigates the potential of electric vehicles (EVs) in a context of a pilot test in Belgium, consisting of car sharing services managed and exploited in small communities. Part of a broader testing activity in the framework of the e-Mobility NSR project, the test had the objective of metering EVs’ charging and consumption in real daily transport operations, but soon it acquired new meanings. It shaped EV sharing services at a small scale, directly managed by the users, which also share energy and maintenance costs. The chosen context was cohousing, a special type of collaborative housing, and four ones were selected in Flanders: two urban units and two larger semi-urban ones. The two urban communities received a prepaid card for reserving and using EVs provided by Cambio, a Belgian car sharing company. The other two cohousings received two EVs and a charging box, organising and running an internal car sharing system for the duration of 1 year. During the tests, quantitative and qualitative data were collected. The paper reports the intermediate results, identifying potential EV sharing consumers, based on their behaviour and attitudes in relation to the condition of the local context in which they live.

Cohousing and EV Sharing: Field Tests in Flanders

Abstract Background In recent decades, cities worldwide are increasingly adopting vehicle access policies and technologies to alleviate the negative externalities related to high car use in the urban built environments. As such, car-oriented infrastructures and mobility policies implemented post-World War II are giving way to sustainable mobility strategies that aim to make cities healthier, more livable and more inclusive for all. Purpose Most of these strategies are considered as stand-alone cases related to a specific environmental, political and social urban context. However, similarities and patterns between different strategies can provide information on the replicability of mobility strategies in other urban contexts. Methods Through a literature review, this paper illustrates the wide range of urban vehicle access regulations (UVAR) applied within sustainable mobility strategies. In addition, we critically examine the process from ideation over design and implementation to operation phase for 12 West-European cities and define what measures are combined to end up with a sustainable mobility strategy. Results This results in a taxonomy of UVAR interventions, subdivided in three categories: spatial interventions, pricing aspects and regulatory measures. We also highlight a number of complementary supportive measures implemented to counter the often restrictive nature of UVAR. Results The paper shows that the strategies highlighted in the case studies are an amalgam of different UVAR and supportive measures. As such, deconstructing various sustainable mobility strategies enables us to shed light on the available UVAR options cities can combine to define consistent and robust sustainable mobility strategies. 

/pdf/sustainable-mobility-strategies-deconstructed-a-taxonomy-of-20nr82k5.pdf

Sidharta Gautama

Papers

Improved classification of urban areas with pixel-based measures of object size

Efficient use of geographical information systems for improving transport mode classification

Describing image content for quality assessment of geospatial data

Cohousing and EV Sharing: Field Tests in Flanders

Sustainable mobility strategies deconstructed: a taxonomy of urban vehicle access regulations