Abstract We propose an extension to the framework of reaction systems such that the environment not only plays the role of contributing entities to the successor state, but also influences the actual generation of products. In particular, we consider that an environment can assign valences to entities, be they used as reactants or inhibitors in a reaction, as well as select the entities which can appear in the successor state as product of a reaction. This provides a framework for immersing reaction system specifications into different contexts. We also derive a notion of polarised reaction systems , where entities can appear only as reactants or inhibitors in any given reaction. We compare the expressive power of the models, also showing some equivalences between them, and test them on three case studies derived from physiology, ecology, and chemistry.

BACKGROUND: Recent years have seen a substantial increase in RNA-seq data production, with this technique becoming the primary approach for gene expression studies across a wide range of non-model organisms. The majority of these organisms lack a well-annotated reference genome to serve as a basis for studying differentially expressed genes (DEGs). As an alternative cost-effective protocol to using a reference genome, the assembly of RNA-seq raw reads is performed to produce what is referred to as a 'de novo transcriptome,' serving as a reference for subsequent DEGs' analysis. This assembly step for conventional DEGs analysis pipelines for non-model organisms is a computationally expensive task. Furthermore, the complexity of the de novo transcriptome assembly workflows poses a challenge for researchers in implementing best-practice techniques and the most recent software versions, particularly when applied to various organisms of interest. RESULTS: To address computational challenges in transcriptomic analyses of non-model organisms, we present HPC-T-Assembly, a tool for de novo transcriptome assembly from RNA-seq data on high-performance computing (HPC) infrastructures. It is designed for straightforward setup via a Web-oriented interface, allowing analysis configuration for several species. Once configuration data is provided, the entire parallel computing software for assembly is automatically generated and can be launched on a supercomputer with a simple command line. Intermediate and final outputs of the assembly pipeline include additional post-processing steps, such as assembly quality control, ORF prediction, and transcript count matrix construction. CONCLUSION: HPC-T-Assembly allows users, through a user-friendly Web-oriented interface, to configure a run for simultaneous assemblies of RNA-seq data from multiple species. The parallel pipeline, launched on HPC infrastructures, significantly reduces computational load and execution times, enabling large-scale transcriptomic and meta-transcriptomics analysis projects.

Abstract Purpose Recent advances in AI, especially in large language models, have created new opportunities to integrate human and artificial agents through shared linguistic capabilities. This paper presents a multi-agent organizational framework in which human agents, LLMs, and specialized agents (narrow AIs) collaborate via dynamic, topic-based group formation. Topic-driven interactions enable agents to coalesce around evolving interests, supported by threshold-based protocols for temporal adaptation, topic emergence, and participation. Methods Within our framework, human agents guide the overall system objectives, while consultant agents (LLMs) provide semantic analysis and mediation, and specialized agents perform focused domain tasks. By leveraging automated topic modeling, the approach eschews rigid ontologies and instead supports adaptive and interpretable content management. Mathematical properties ensure system coherence—across roles, tasks, and timescales—while allowing the natural evolution of interests and groups. Results We illustrate the framework’s versatility with example scenarios in emergency response, healthcare research, and financial decision-making, emphasizing how human decision-makers, LLM-based consultants, and specialized worker agents jointly fulfill complex goals through transparent topic alignment and threshold-driven coordination. This formalization advances human-computer interaction as a multi-agent phenomenon that integrates human insight with the strengths of next-generation AI models in a cohesive, evolving system.

Abstract We proceed in the investigation of models of environment in reaction systems, beyond the simple models establishing context-independent or arbitrary processes, to consider restrictions on both the sets of entities that an environment can contribute to a reaction system and the set of entities that a reaction system can produce. Such a distinction is reminiscent of that between self and non-self in living systems and is examined in this paper under different respects. We discuss applications of the model to living systems and security, as well as relations between families of reaction systems with self distinction, unveiling peculiarities of processes with self distinction with respect to the classical notion, for both context-independent processes and processes with arbitrary context. In particular, we introduce a notion of simulation, through which we show that the restriction to systems with self distinction does not entail any loss in the set of realisable behaviours, and that there exists a family of reaction systems with self distinction able to simulate context-independent behaviours of classical reaction systems through processes with arbitrary context. Moreover, using 2-step simulations, we show that every reaction system with self distinction can be reduced to a system (with self-distinction) using only reactions where no set of reactants and inhibitors mixes entities from both the self and the non-self sets.

Abstract Modelling languages are intensively used in paradigms like model-driven engineering to automate all tasks of the development process. These languages may have variants, in which case the need arises to deal with language families rather than with individual languages. However, specifying the syntax and semantics of each language variant separately in an enumerative way is costly, hinders reuse across variants, and may yield inconsistent semantics between variants. Hence, we propose a novel, modular and compositional approach to describing product lines of modelling languages. It enables the incremental definition of language families by means of modules comprising meta-model fragments, graph transformation rules, and rule extensions. Language variants are configured by selecting the desired modules, which entails the composition of a language meta-model and a set of rules defining its semantics. This paper describes: a theory for checking well-formedness, instantiability, and consistent semantics of all languages within the family; an implementation as an Eclipse plugin; and an evaluation reporting drastic specification size and analysis time reduction in comparison to an enumerative approach.

BACKGROUND: The availability of transcriptomic data for species without a reference genome enables the construction of de novo transcriptome assemblies as alternative reference resources from RNA-Seq data. A transcriptome provides direct information about a species' protein-coding genes under specific experimental conditions. The de novo assembly process produces a unigenes file in FASTA format, subsequently targeted for the annotation. Homology-based annotation, a method to infer the function of sequences by estimating similarity with other sequences in a reference database, is a computationally demanding procedure. RESULTS: To mitigate the computational burden, we introduce HPC-T-Annotator, a tool for de novo transcriptome homology annotation on high performance computing (HPC) infrastructures, designed for straightforward configuration via a Web interface. Once the configuration data are given, the entire parallel computing software for annotation is automatically generated and can be launched on a supercomputer using a simple command line. The output data can then be easily viewed using post-processing utilities in the form of Python notebooks integrated in the proposed software. CONCLUSIONS: HPC-T-Annotator expedites homology-based annotation in de novo transcriptome assemblies. Its efficient parallelization strategy on HPC infrastructures significantly reduces computational load and execution times, enabling large-scale transcriptome analysis and comparison projects, while its intuitive graphical interface extends accessibility to users without IT skills.

Smart contracts show a high potential to make supply chain management strategies epochally leap towards higher levels of productivity, not only in the functioning of production processes but also in terms of product innovation and overall economic returns. This article illustrates the principle of Income Sharing as a highly performing economic strategy for supply chains with a natural implementation in blockchain smart contracts. It proposes a blockchain-based architecture that uses smart contracts to implement various algorithmic versions of the Income Sharing principle among companies participating in a supply chain. The formation of the total income and its consequent redistribution are calculated taking into account the role of the technological platform automating these procedures, which therefore becomes a party to the inter-company business project of a supply chain in the alternative roles, as feasible in business practice, of Blockchain-as-a-Service and Blockchain-as-a-Partner. The approach is implemented on Hyperledger Fabric, the most widespread platform for private and consortium blockchains. We compare and justify this design choice with the alternative given by public blockchains, with specific attention to Ethereum.

Modelling languages are intensively used in paradigms like model-driven engineering to automate all tasks of the development process. These languages may have variants, in which case the need arises to deal with language families rather than with individual languages. However, specifying the syntax and semantics of each language variant separately is costly, hinders reuse across variants, and may yield inconsistent semantics between variants.

A key component of blockchain technology is the ledger, viz., a database that, unlike standard databases, keeps in memory the complete history of past transactions as in a notarial archive for the benefit of any future test. In second-generation blockchains such as Ethereum, the ledger is coupled with smart contracts, which enable the automation of transactions associated with agreements between the parties of a financial or commercial nature. The coupling of smart contracts and ledgers provides the technological background for very innovative application areas, such as Decentralized Autonomous Organizations (DAOs), Initial Coin Offerings (ICOs), and Decentralized Finance (DeFi), which propelled blockchains beyond cryptocurrencies that were the only focus of first generation blockchains, such as Bitcoin. However, the currently used implementation of smart contracts as arbitrary programming constructs has made them susceptible to dangerous bugs that can be exploited maliciously and has moved their semantics away from that of legal contracts. We propose here to recompose the split and recover the reliability of databases by formalizing a notion of contract modeled as a finite-state automaton with well-defined computational characteristics derived from encoding in terms of allocations of resources to actors, as an alternative to the approach based on programming. To complete the work, we use temporal logic as the basis for an abstract query language that is effectively suited to the historical nature of the information kept in the ledger.

Private distributed ledgers (DLs) appear to be more successful than public blockchains at supporting innovative business ecosystems in the form of enterprise consortia. In fact, by untying transactions from the native cryptocurrencies of public blockchains, consortium DLs can smoothly interface with the various sectors of the traditional economy with the capability to innovate it. Furthermore, eliminating mining greatly increases the computational efficiency of these technologies compared to public blockchains. So far, consortium DLs have been used primarily for traceability and document management. Now the time is ripe to move forward, impacting business and profitability. We describe here a consortium model based on distributed ledger technology that applies the old principle that “unity is strength” by deploying the economic method of Revenue Sharing as a smart contract. We illustrate the model through a prototype implementation on Hyperledger Fabric, the most widely adopted among the platforms for consortium distributed ledgers.

We present an approach to integrating extemporary annotations on topics of interest to a user with information on the context in which the annotation is being taken. Context is here defined in terms of the set of surrounding devices (Bluetooth Low Energy enabled smartphones, WiFi hotspots) and the current calendar event. Contextify is a context-aware Android application which detects the current context and organizes the notes taken within context, allowing a form of context-based retrieval. Thus, the detected context represents location, other people present (referring to their BLE equipped smartphones as user proxies), and events. Users can easily retrieve notes when they return to the context where they created them. A context similarity algorithm derived from the Jaro-Winkler string similarity algorithm is used to compare contexts. Each note is tagged by the user and the system suggests the most appropriate tag, among the already used ones, at annotation creation time: the suggestion is based on the similarity of the current context with the contexts associated with previously tagged notes.

With technologies for massively parallel genome sequencing available, bioinformatics has entered the “big data” era. Developing applications in this field involves collaboration of domain experts with IT specialists to specify programs able to query several sources, obtain data in several formats, search them for significant patterns and present the obtained results according to several types of visualisation. Based on the experience gained in developing several Web portals for accessing and querying genomics and proteomics databases, we have derived a meta-model of such portals and implemented BioWebEngine, a generation environment where a user is assisted in specifying and deploying the intended portal according to the meta-model.

The study of algebraic modelling of labelled non-deterministic concurrent processes leads us to consider a categoryLB, obtained from a complete meet-semilatticeBand fromB-valued equivalence relations. We prove that, ifBhas enough properties, thenLBpresents a two-fold internal logical structure, induced by two doctrines definable on it: one related to its families of subobjects and one to its families of regular subobjects. The first doctrine is Heyting and makesLBa Heyting category, the second one is Boolean. We will see that the difference between these two logical structures, namely the different behaviour of the negation operator, can be interpreted in terms of a distinction between non-deterministic and deterministic behaviours of agents able to perform computations in the context of the same process. Moreover, the sorted first-order logic naturally associated withLBcan be extended to a modal/temporal logic, again using the doctrinal setting. Relations are also drawn to other computational models.

We discuss the design and development of HCI models for authentication based on gait and gesture that can be supported by mobile and wearable equipment. The paper proposes to use such biometric behavioral traits for partially transparent and continuous authentication by means of behavioral patterns.

Model-based development relies on models in different phases for different purposes, with modelling patterns being used to document and gather knowledge about good practices in specific domains, to analyse the quality of existing designs, and to guide the construction and refactoring of models. Providing a formal basis for the use of patterns would also support their integration with existing approaches to model transformation. To this end, we turn to the commonly used, in this context, machinery of graph transformations and provide an algebraic-categorical formalization of modelling patterns, which can express variability and required/forbidden application contexts. This allows the definition of transformation rules having patterns in left and right-hand sides, which can be used to express refactorings towards patterns, change the use of one pattern by a different one, or switch between pattern variants. A key element in our proposal is the use of operations to abstract models into patterns, so that they can be manipulated by pattern rules, thus leading to a rewriting mechanism for classes of graphs described by patterns and not just individual graphs. The proposal is illustrated with examples in object-oriented software design patterns and enterprise architecture patterns, but can be applied to any other domain where patterns are used for modelling.

We describe an ongoing co-design project for a collaborative distributed space, based on multi-touch technologies, through which older adults can explore self-produced and on-line multimedia resources for the production of short video clips. The project's main goal is to foster the active participation of older people, already experiencing collaboration in daily activities in care centres, as producers of content related to their experience and know-how and to the activities carried out in the centres, to be shared in a community context. An interface, based on an interactive table, some tablets and a television set, is sketched.

The recently introduced notion of Modeling Spider Diagrams immerses Spider Diagrams in a modelling context. We enhance here their capabilities by including references to resources that are required to persist in some state or to perform some state transition. To this end, we introduce a notion of anno-tation, which can be expressed through a restriction of Conceptual Spider Diagrams (extended over domains), where domains are partially ordered and predicate edges have to be consistent with this ordering. The resulting notion of Annotated Modeling Spider Diagrams can then be used to express constraints on the usage of resources, so that conformance to a policy can be verified with respect to resource availability. We provide additional options for user-modelling choices via a natural equivalent representation in terms of (classical) Spider Diagrams, restricted on the domain of interest, and the use of colour for domain identification.

We extend the study of networks of evolutionary processors accepting words to a similar model, processing rectangular pictures. To this aim, we introduce accepting networks of evolutionary picture processors and investigate their computational power. We show that these networks can accept the complement of any local picture language as well as picture languages that are not recognizable. Some open problems regarding decidability issues and closure properties are finally discussed.

Euler diagrams (EDs) are a popular means of visualizing set-based relationships. We describe the advancement of EulerSketch, based on a novel encoding of EDs. In particular, EulerSketch allows a user to sketch and interact with EDs and automatically produce a code representation for them. It also enables code storage and retrieval and permits to automatically re-draw an ED from its code representation. Some features are inspired by earlier works [4, 6, 3].

We propose a framework for the construction of collaborative services, based on a technique for coding content generated both by authoritative organisations and by users on interactive supports. In particular, we show how users can interact in an enriched and focused way with paper documents and object labels, both in desktop and mobile settings. We argue that combining immediate access to online information from physical support with creation and retrieval of annotations, while keeping in focus the context of their creation, would create a collaboration space by extending the ways of interacting with documents and objects and the way in which these allow seamless access to both centralised and distributed information.

Timed Modelling Spider Diagrams (TMSDs) are a visual language which supports the modeling of object-oriented systems with time constraints. They are used to define policies in which TMSDs specify admissible evolutions of the state of some instance. We define a process for deriving a rewriting system from a policy specification, so that the generated system defines a language of sequences of basic TMSDs satisfying the policy. Moreover, by identifying the different ways in which the constraints set by the policy can be violated, we can produce special rules whose application results in erroneous sequences. The resulting transformation systems can be used both to simulate possible behaviours when reasoning on the definition of policies and to test policy verification algorithms.

We present SWift (SignWriting improved fast transcriber), an advanced editor for computer-aided writing and transcribing using SignWriting (SW). SW is devised to allow deaf people and linguists alike to exploit an easy-to-grasp written form of (any) sign language. Similarly, SWift has been developed for everyone who masters SW, and is not exclusively deaf-oriented. Using SWift, it is possible to compose and save any sign, using elementary components called glyphs. A guided procedure facilitates the composition process. SWift is aimed at helping to break down the "electronic" barriers that keep the deaf community away from Information and Communication Technology (ICT). The editor has been developed modularly and can be integrated everywhere the use of SW, as an alternative to written vocal language, may be advisable.

In the proposed demo, a new approach to e-learning environments for deaf people is illustrated. Using a fully iconic web-based environment, a tutor can define, generate and test e-learning courses for deaf people, which are automatically managed, published and served by the system itself.

Spider Diagrams are a well-established visual language used to specify sets, their relationships, and constraints on their cardinalities but they have no means of specifying temporal aspects of a system. Timed Spider Diagrams are an evolution enabling the specification of temporal constraints, with a granular-based time system, for use in areas such as policy specification. In this paper we introduce event based actions to this framework and illustrate with examples from models of an automatic parking meter system, and an internet billing system.

MADCOW 2.0 is a system for annotation of Web content, supporting the production and exploration of personal and public annotations on text, images and videos in a Web page. Its design starts from the main requirement that the annotation activity does not have to disrupt the normal browsing of Web pages by a user. MADCOW 2.0 allows interaction with the annotated portions of the page to provide access to the annotation content. Conversely, the representation of the existing notes supports different forms of exploration of the Web page, and can become the starting point for further navigation over the Web. A uniform style of interaction has been adopted for creating and accessing annotations on text, images and videos, and some novel solutions have been introduced to cope with overlaps between the annotated portions. The annotation user experience is facilitated by enabling forms of in-place annotation and manipulation of both the annotated portion and the annotation content.

While interaction patterns are becoming widespread in the field of interface design, their definitions do not enjoy a common standard yet, as is for software patterns. Moreover, patterns are developed for diverse design aspects, reflecting the complexity of the field. As a consequence, research on formalization of interaction patterns is not developed, and few attempts have been made to extend techniques developed for design pattern formalization. We show here how an extension to our recent approach to pattern formalization can be usefully employed to formalize some classes of interaction patterns, to express relations among them, and to detect conflicts.

Spatial structures are particularly suited to the definition of parallel programs, due to their homogeneity. The Movie-based Programming Framework allows specification of computations on regular networks of processors, and the visualization of the computation progress as processors are activated. Computations over spatial structures are specified by composing independent views on propagation of control flows and formulae defining local computations. A shape pattern indicates which processors have to be active during a specific phase of the computation. A visit pattern defines the law of propagation for actual processor activation. By combining these types of patterns, we achieve sophisticated forms of specification. In particular, one specifies visitors implementing collective communication schemas widely used in parallel programming: broadcast, gather, scatter and reduction. As a result, automatic generation of visit algorithms adapted to different network configurations is made possible, thus facilitating experimentation with different laws and their visualization.

We present a new approach - based on graph transformation - to incremental specification of the operational (execution) semantics of visual languages. The approach combines editing rules with two meta-models: one to define the concrete syntax and one for the static semantics. We introduce the notion of action patterns, defining basic actions (e.g. consuming or producing a token in transition-based semantics), in a way similar to graph transformation rules. The application of action patterns to a static semantics editing rule produces a meta-rule, to be paired with the firing of the corresponding syntactic rule to incrementally build an execution rule. An execution rule is thus tailored to any active element (e.g. a transition in a Petri net model) in the model. Examples from Petri nets, state automata and workflow languages illustrate these ideas.

Current research in interaction aims at defining new types of multimedia and multimodal experience, at enriching everyday objects and environments with the ability to capture user actions and intentions, and at integrating real and virtual sources of information, typically exploiting the visual channel. These forms of interaction usually require dedicated architectures, often relying on different component models, and with rigid types of configuration. We present an approach to the integration of real and virtual world sensors and effectors, and of traditional multimedia environments within a single component-based architecture. Environments in this architecture are defined as networks of plugins, each equipped with computational, presentation and communication capabilities. Examples of integrated environments produceable with this architecture are given.

The use of the Web and the diffusion of knowledge management systems makes it possible to base discussions upon a vast set of documents, many of which also include links to multimedia material, such as images or videos. This perspective could be exploited by allowing a team to collaborate by exchanging and retrieving annotated multimedia documents (text, images, audio and video). We designed and developed a digital annotation system, MADCOW, to assist users in constructing, disseminating, and retrieving multimedia annotations of documents, supporting collaborative activities to build a web of decision-related documents. We made a strong effort in designing the user interface and we tested it with 24 users. We describe a scenario in which annotation plays a crucial role, where the object of the collaboration is a politically and artistically important palace of Rome, for which the availability of images and historical documentation is fundamental in order to take informed decisions. We demonstrate the MADCOW interface and its use in the restoration team. The annotations can be used to support teamwork as well as to offer the public some reasoned integration and guide to the available material.

We present HYPERSCORE, a system for interactive execution of musical performances coupled with the generation system CLIPSCORE, in which a performer can virtually manipulate execution parameters through gestural interaction in real time. HYPERSCORE allows users to exploit their perceptive capabilities in terms of space and time in order to control and manipulate, via a multimodal interaction, the main component of music, namely time.

Visual interaction involves different types of resources, both at the computational and at the human end of the interaction loop. The roles played by resources, with respect to the development of interactive tasks, are also discussed

The Syntax Manager (SynM) for interactive construction of visual rewriting systems allows users to specify visual languages, by compounding the definition of concrete syntax with metamodel information on the role alphabet elements can play. A concrete rule definition is constrained to comply with the admissible relations for the roles of the elements in it. Hybrid elements can play different roles with respect to different relations.

We propose a resource-based model of human-computer interaction, and discuss how WIPPOG, a rule-based language for multiset transformations, can support such a model. WIPPOG is here adopted as an enactment mechanism for task models expressed by a visual notation. We exploit this resource-based perspective to model some common interaction errors and suggest possible prevention and mitigation strategies.

Digital annotation of multimedia documents adds information to a document (e.g. a web page) or parts of it (a multimedia object such as an image or a video stream contained in the document). Digital annotations can be kept private or shared among different users over the internet, allowing discussions and cooperative work. We study the possibility of annotating multimedia documents with objects which are in turn of multimedial nature. Annotations can refer to whole documents or single portions thereof, as usual, but also to multi-objects, i.e. groups of objects contained in a single document. We designed and developed a new digital annotation system organized in a client-server architecture, where the client is a plug-in for a standard web browser and the servers are repositories of annotations to which different clients can login. Annotations can be retrieved and filtered, and one can choose different annotation servers for a document. We present a platform-independent design for such a system, and illustrate a specific implementation for Microsoft Internet Explorer on the client side and on JSP/MySQL for the server side.

Metamodeling frameworks for the definition and management of visual languages allow the implementation of visual environments based on some abstract notion of visual entity and of relations among them. We propose a suite of metamodels able to accommodate most commonly used visual paradigms, built as progressive specialisation of a root meta-meta model. This approach facilitates the design and implementation of new, general purpose as well as domain specific, visual languages by allowing the progressive construction of language-independent service layers.

Visual interaction processes are modeled in this paper as sequences of visual sentences in which for each visual sentence only a limited set of user actions is possible. We introduce the notion of "dynamic visual language" as a weakly ordered set of visual sentences characterized by the presence of common elements. We present a formal model of derivation of visual sentences in a dynamic visual language in which each visual sentence specifies the possible actions which can be performed on it and the possible transformations it can go through. In this way, we offer a formal setting in which the interaction process can be formally specified. A user interface can be derived from the formal specification, so that it embeds proper context elements which limit user disorientation. The concepts are illustrated by the user interaction with a prototype of a digital library developed at the University of Bari.

We explore interaction as a basis for human-computer comprehension. Perceptual experience is organised through categories establishing the kind of distinctions imposable on perceived phenomena. An interactive tool exploiting a similar categorisation in graphics is integrated into a system for subjective retrieval of images, so that the user interacts with active regions supporting the same type of distinction.

We continue the study of pictorial languages as formalized in Ref. 1 (based on the operations of shifting and superposing elementary pictures). If the pixels are superposed without composing their colors, then we produce only recursive languages. When the colors of the superposed pixels can be composed, then any array grammar can be simulated (hence, all recursively enumerable languages can be obtained). Bidimensional pictorial frameworks with a nonrecursive membership problem are obtained in the restricted case when (1) we do not allow the superposition of nontransparent pixels, excepting the fact that (2) for each color there is a complementary color, which, superposed on the original one, leads to a transparent pixel.

Computers are increasingly being seen not only as computing tools but more so as communication tools, thus placing special emphasis on human-computer interaction (HCI). In this article, the focus is on visual HCI, where the messages exchanged between human and computer are images appearing on the computer screen, as usual in current popular user interfaces. We formalize interactive sessions of a human-computer dialogue as a structured set of legal visual sentences, i.e., as a visual language, and show how rewriting systems can be generalized to specify both the pictorial and the computational aspects of visual languages. To this end, Visual Conditional Attributed Rewriting (VCARW) systems are introduced, and use for specification of visual languages. These specifications are given as inputs to a procedure illustrated in the article as a system of algorithms, which automatically generates control mechanisms of the interaction, thus favoring the design of more reliable and usable systems.

A visual environment for end-user visual programming and computing with populations of agents is presented.

The paper addresses the problem of controlling situated image understanding processes. Two complementary control styles are considered and applied cooperatively, a deliberative one and a reactive one. The role of deliberative control is to account for the unpredictability of situations, by dynamically determining which strategies to pursue, based on the results obtained so far and more generally on the state of the understanding process. The role of reactive control is to account for the variability of local properties of the image by tuning operations to subimages, each one being homogeneous with respect to a given operation. A variable organization of agents is studied to face this variability. The two control modes are integrated into a unified formalism describing segmentation and interpretation activities. A feedback from high level interpretation tasks to low level segmentation tasks thus becomes possible and is exploited to recover wrong segmentations. Preliminary results in the field of liver biopsy image understanding are shown to demonstrate the potential of the approach.

A novel definition of visual languages allows a uniform approach to satisfying the needs of visual reasoning faced in visual human-computer interaction. The way the machine associates a computational meaning with an image, and conversely, the way it generates an image on the screen from a computation are formally described. A definition of a visual sentence and of a visual language as a set of visual sentences is discussed. A hierarchy of visual languages is derived in relation to the requirements for intelligible, manageable and trustable interaction between humans and computers.

Two complementary control styles, deliberative and reactive types, are applied to account for unpredictability of situations and variability of local properties in image interpretation processes. Deliberative control is based on attributed grammars and local variability is managed by a variable organisation of agents. Feedback from interpretation to segmentation becomes possible and is exploited to recover wrong segmentation. Experimental results illustrate the potential of the approach.

The paper introduces the concepts of image interpretation by form feature and interpretation strategy as a fall out of observing human interpreters at work. It discusses the need for flexible control in an automatic interpreter, outlining an automatic interpretation process and the required meta (or reflective) mechanisms.>

A method for identifying and counting cells and biological structures in histological liver tissue sections is presented. The method uses several image processing operators. The final classification results from contextual fusion of different clues. The overall flow of information among the active components of the cell description procedure (CDP) is outlined. The CDP is constructed and validated by a sequence of exploration and confirmation phases. Exploration takes place when a set of known images that is progressively improved until the results are satisfactory is worked out through the CDP. The confirmation phase may then start, working out a new set of images (of the same type but not analyzed before) and proceeding until acceptable sets of agents are generated. The completion of a first exploration phase has identified some major fields of improvement such as the identification of revision criteria for the refinement of the F-nuclei class.>

An automatic system which sets up high-level descriptions of the ECG, based on its medical shape features, from the low-level descriptions of the numerical techniques is discussed. These descriptions are organized into a description scheme, thus making it possible for physicians to use visual interpretation to control the whole description process. At each moment during an interpretation activity, the content of the data structure can be shown to the user in a graphical manner. Using this method, the progress of the process and the reasons why an interpretation strategy is followed can be controlled by the user.>