Leen Kiat Soh

Texture analysis of SAR sea ice imagery using gray level co-occurrence matrices

This paper presents a preliminary study for mapping sea ice patterns (texture) with 100-m ERS-1 synthetic aperture radar (SAR) imagery. The authors used gray-level co-occurrence matrices (GLCM) to quantitatively evaluate textural parameters and representations and to determine which parameter values and representations are best for mapping sea ice texture. They conducted experiments on the quantization levels of the image and the displacement and orientation values of the GLCM by examining the effects textural descriptors such as entropy have in the representation of different sea ice textures. They showed that a complete gray-level representation of the image is not necessary for texture mapping, an eight-level quantization representation is undesirable for textural representation, and the displacement factor in texture measurements is more important than orientation. In addition, they developed three GLCM implementations and evaluated them by a supervised Bayesian classifier on sea ice textural contexts. This experiment concludes that the best GLCM implementation in representing sea ice texture is one that utilizes a range of displacement values such that both microtextures and macrotextures of sea ice can be adequately captured. These findings define the quantization, displacement, and orientation values that are the best for SAR sea ice texture analysis using GLCM.

A comprehensive, automated approach to determining sea ice thickness from SAR data

Documents an approach to sea ice classification through a combination of methods, both algorithmic and heuristic. The resulting system is a comprehensive technique, which uses dynamic local thresholding as a classification basis and then supplements that initial classification using heuristic geophysical knowledge organized in expert systems. The dynamic local thresholding method allows separation of the ice into thickness classes based on local intensity distributions. Because it utilizes the data within each image, it can adapt to varying ice thickness intensities to regional and seasonal changes and is not subject to limitations caused by using predefined parameters. >

ARKTOS: an intelligent system for SAR sea ice image classification

We present an intelligent system for satellite sea ice image analysis named Advanced Reasoning using Knowledge for Typing Of Sea ice (ARKTOS). ARKTOS performs fully automated analysis of synthetic aperture radar (SAR) sea ice images by mimicking the reasoning process of sea ice experts. ARKTOS automatically segments a SAR image of sea ice, generates descriptors for the segments of the image, and then uses expert system rules to classify these sea ice features. ARKTOS also utilizes multisource data fusion to improve classification and performs belief handling using Dempster-Shafer. As a software package, ARKTOS comprises components in image processing, rule-based classification, multisource data fusion, and graphical user interface-based knowledge engineering and modification. As a research project over the past ten years, ARKTOS has undergone phases such as knowledge acquisition, prototyping, refinement, evaluation, deployment, and operationalization at the U.S. National Ice Center. In this paper, we focus on the methodology, evaluations, and classification results of ARKTOS.

A Real-Time Negotiation Model and A Multi-Agent Sensor Network Implementation

This paper describes a negotiation model that incorporates real-time issues for autonomous agents. This model consists of two important ideas: a real-time logical negotiation protocol and a case-based negotiation model. The protocol integrates a real-time Belief-Desire-Intention (BDI) model, a temporal logic model, and communicative acts for negotiation. This protocol explicitly defines the logical and temporal relationships of different knowledge states, facilitating real-time designs such as multi-threaded processing, state profiling and updating, and a set of real-time enabling functional predicates in our implementation. To further support the protocol, we use a case-based reasoning model for negotiation strategy selection. An agent learns from its past experience by deriving a negotiation strategy from the most similar and useful case to its current situation. Guided by the strategy, the agent negotiates with its partners using an argumentation-based negotiation protocol. The model is time and situation aware such that each agent changes its negotiation behavior according to the progress and status of the ongoing negotiation and its current agent profile. We apply the negotiation model to a resource allocation problem and obtain promising results.

Segmentation of satellite imagery of natural scenes using data mining

The authors describe a segmentation technique that integrates traditional image processing algorithms with techniques adapted from knowledge discovery in databases (KDD) and data mining to analyze and segment unstructured satellite images of natural scenes. They have divided their segmentation task into three major steps. First, an initial segmentation is achieved using dynamic local thresholding, producing a set of regions. Then, spectral, spatial, and textural features for each region are generated from the thresholded image. Finally, given these features as attributes, an unsupervised machine learning methodology called conceptual clustering is used to cluster the regions found in the image into N classes-thus, determining the number of classes in the image automatically. They have applied the technique successfully to ERS-1 synthetic aperture radar (SAR). Landsat thematic mapper (TM), and NOAA advanced very high resolution radiometer (AVHRR) data of natural scenes.

Concept inventories in computer science for the topic discrete mathematics

This report describes concept inventories, specialized assessment instruments that enable educational researchers to investigate student (mis)understandings of concepts in a particular domain. While students experience a concept inventory as a set of multiple-choice items taken as a test, this belies its purpose, its careful development, and its validation. A concept inventory is not intended to be a comprehensive instrument, but rather a tool that probes student comprehension of a carefully selected subset of concepts that give rise to the most common and pervasive mismodelings. The report explains how concept inventories have been developed and used in other STEM fields, then outlines a project to explore the feasibility of concept inventories in the computing field. We use the domain of discrete mathematics to illustrate a suggested plan of action.

Adaptive, Confidence-Based Multiagent Negotiation Strategy

We propose an adaptive 1-to-many negotiation strategy for multiagent coalition formation in dynamic, uncertain, real-time, and noisy environments. Our strategy focuses on multi-issue negotiations where each issue is a request from the initiating agent to the responding agent. The initiating agent conducts multiple concurrent negotiations with responding agents and in each negotiation it employs (1) a pipelined, one-at-a-time approach, or (2) a confidence-based, packaged approach. In the former, lacking knowledge on the responding agent, it negotiates one issue at a time. In the latter, with confident knowledge of the past behavior of the responding agent, it packages multiple issues into the negotiation. We incorporate this adaptive strategy into a multi-phase coalition formation model (MPCF) in which agents learn to form coalitions and perform global tasks. The MPCF model consists of three phases: coalition planning, coalition instantiation and coalition evaluation. In this paper, we focus on the instantiation phase where the negotiations take place.

ClassroomWiki: A Collaborative Wiki for Instructional Use with Multiagent Group Formation

Wikis today are being used as a tool to conduct collaborative writing assignments in classrooms. However, typical Wikis do not adequately address the assessment of individual student contributions toward their groups or provide any automated group formation mechanism. To improve these aspects, we have designed and implemented ClassroomWiki - a web-based collaborative Wiki writing tool. For the students, ClassroomWiki provides a web interface for writing and revising their groups Wiki and a topic-based forum for discussing their ideas during collaboration. When the students collaborate, ClassroomWiki tracks all student activities and builds detailed student models that represent their contributions toward their groups. For the teacher, ClassroomWiki provides a multiagent framework that uses the student models to form student groups to improve the collaborative learning of students. To investigate the impact of ClassroomWiki, we have conducted a three-week-long collaborative Wiki writing assignment in a university-level history course. The results suggest that ClassroomWiki can 1) improve the collaborative learning outcome of the students by its group formation framework, 2) help the teacher better assess a students contribution toward his or her group and avoid free riding, and 3) facilitate specific and precise teacher intervention with accurate and detailed tracking of student activities.

Satisficing coalition formation among agents

In a multiagent system where each agent has only an incomplete view of the world, optimal coalition formation is difficult. Coupling that with real-time and resource constraints often makes the rationalization process infeasible or costly. We propose a coalition formation approach that identifies and builds sub-optimal yet satisficing coalitions among agents to solve a problem detected in the environment. All agents are peers and autonomous. Each is motivated to conserve its own resources while cooperating with other agents to achieve a global task or resource allocation goal. The (initiating) agent-that detects a problem-hastily forms an initial coalition by selecting neighboring agents that it considers to have high potential utilities, based on the capability of each neighbor and its respective inter-agent relationships. The initiating agent next finalizes the coalition via multiple concurrent 1-to-1 negotiations with only neighbors of high potential utility, during which constraints and commitments are exchanged in an argumentation setting.

A dissimilarity function for clustering geospatial polygons

The traditional point-based clustering algorithms when applied to geospatial polygons may produce clusters that are spatially disjoint due to their inability to consider various types of spatial relationships between polygons. In this paper, we propose to represent geospatial polygons as sets of spatial and non-spatial attributes. By representing a polygon as a set of spatial and non-spatial attributes we are able to take into account all the properties of a polygon (such as structural, topological and directional) that were ignored while using point-based representation of polygons, and that aid in the formation of high quality clusters. Based on this framework we propose a dissimilarity function that can be plugged into common state-of-the-art spatial clustering algorithms. The result is clusters of polygons that are more compact in terms of cluster validity and spatial contiguity. We show the effectiveness and robustness of our approach by applying our dissimilarity function on the traditional k-means clustering algorithm and testing it on a watershed dataset.