Application-specific Knowledge Research Articles

Abstract 3523: Self-supervised foundation model captures high-dimensional morphology data from single cell brightfield images

Abstract Our knowledge of cancer biology has advanced with the characterization of distinct cell types and cell states within heterogeneous tumor environments. Many methods for imaging and sorting tumor cells require biomarker labels that alter cell characteristics and create a selection bias. The use of label-free single-cell methods should further improve cancer studies involving viable, unperturbed cells for downstream assays such as RNA-Seq, cell culture expansion, and functional assays. To eliminate biomarkers and enable broader assessment of cells, we developed REM-I, a platform that characterizes and sorts unlabeled single cells based on high-dimensional morphology. Cells are captured with brightfield imaging and processed in real-time by self-supervised deep learning models to generate quantitative AI embeddings representative of cell morphology. A significant technical challenge in building REM-I was developing an AI model based on extracted features from cell images without prior knowledge of cell types, cell preparation, or other application-specific knowledge. Accordingly, we developed the Human Foundation Model (HFM), a hybrid architecture that combines self-supervised learning (SSL) and morphometrics (computer vision) to extract 115 dimensional embeddings representing cell morphology from high-resolution REM-I cell images. SSL produces a foundation model with high generalization capacity that enables hypothesis-free sample exploration and efficient generation of application-specific models. Meanwhile, computer vision extracts features that represent measurable and interpretable concepts (e.g., cell size, shape, texture, intensity). The training process for the HFM self-supervised backbone model utilizes the discriminatory power of supervised tasks. Using synthetic cells and three cancer cell lines, we trained and then validated the reproducibility and generalization capabilities of the resulting model. Results show combining deep learning and morphometrics improve interpretability of data and enable rapid characterization and classification of tumor cells with high accuracy. We also report on the Deepcell® Axon data suite, a tool to analyze data and customize reusable workflows. This includes the ability to store and manage data, visualize high dimensional data as low-dimensional projections, and train classifiers to identify and sort cell populations. To enable compatibility with user-preferred downstream analysis pipelines, Axon provides data export options for images, plots, and embeddings. Our approach allows users of all skill levels to access and interpret AI-enabled morphological profiling. Applications of REM-I include hypothesis-free evaluation of heterogeneous tumor samples, label-free cancer cell enrichment, characterization of distinct cell states, and multi-omic integration. Citation Format: Kiran Saini, Senzeyu Zhang, Ryan Carelli, Kevin B. Jacobs, Amy Wong-Thai, Cris Luengo, Vivian Lu, Anastasia Mavropoulos, Andreja Jovic, Jeanette Mei, Thomas Vollbrecht, Stephane C. Boutet, Mahyar Salek, Maddison Masaeli. Self-supervised foundation model captures high-dimensional morphology data from single cell brightfield images [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 3523.

Mjolnir: Breaking the Glass in a Publicly Verifiable Yet Private Manner

This paper formally investigates the problem of unauthorized yet required access to electronically protected information, a.k.a. Break-the-Glass (BtG) access. Reflecting on the rising deployment of such protocols in the current digitized healthcare system, we present Mjolnir1, a blockchain-based BtG framework that offers accountability of unauthorized accesses by healthcare practitioners, dependable right of notification to patients, and privacy of healthcare records accesses. Mjolnir is a smart contract-based protocol which provides undisputed public verifiability of the identity of BtG access entities while maintaining their anonymity except from concerned individual patients, hence protecting the patients’ privacy. We employ an application specific non-interactive cryptographic zero knowledge proof system which ensures that the signing entity (healthcare practitioner) belongs to a given authorized group and that the anonymity of their identity is only revocable by a given opening entity (patient). The security of our system relies on the hardness of the discrete logarithm and decisional Diffie-Hellman problems in elliptic curve groups, and the utilized proof system requires no trusted setup. We formally define and prove the security goals of Mjolnir, provide a proof of concept blockchain implementation on Ethereum, and report on performance experiments and comparisons with other generic zero knowledge proof systems.

A reference implementation for knowledge assisted robot development for planetary and orbital robotics

The use of modular, yet reusable and interchangeable components offers flexibility for system designers and thus leads to a higher degree of reconfigurability with cost effective, adaptable and scalable solutions. The development of a robotic system remains, however, a complex task with increasing demands depending on the targeted capabilities, functionalities and operational constraints of the robot. This requires expert knowledge in mechanical, electrical and software engineering, as well as tools that can collect, process and use relevant data. As a result, the amount of available and managed information is increasing, while its relevance, usability, and knowledge that can be extracted from it are consequently decreasing: a situation described as information overload paradox.The design of a modular system can be based on the use of standardized modules with universal interconnects, but it should be aligned with a strong formalism, e.g. the semantic representation of components, to allow for a better management of the overall complexity. An ontology-based knowledge representation, for instance, which provides domain and application specific knowledge for the design of robotic systems, can be used as a method of storing and sharing data in a uniform, machine-readable and standardized way.In this paper, we introduce the Knowledge-based Open Robot voCabulary as Utility Toolkit (korcut) as a core component to support the ontology-driven development of robotic systems as part of a reference implementation of the Q-Rock development cycle. Q-Rock gathers methods to assist users in designing robot configuration by: (a) automatically exploring robot capabilities based on robot hardware, (b) proposing robot designs to meet the user’s needs, (c) refining the proposed robot designs. We use korcut to improve the robot design process in orbital and planetary robotics from the requirements definition to the development phases of complex modular robotic structures (e.g. composition of an unmanned ground vehicle with robotic arms). This is implemented by embedding semantic component descriptions in a state-of-the-art open-source 3D modeling software. Furthermore, the korcut ontology family includes various sub-ontologies to address specific space-related tasks, such as the modeling of a Standard Interconnect (SI) for On Orbit Services (OOS) and Orbital Factory. The ontology design follows criteria that have been derived from a survey conducted as part of this work. This paper also covers an evaluation of its current applications from a methodological, knowledge representation, and software tool perspectives. In the application part, we introduce the use of the ontology to model a multi-functional SI that enables mechanical connections between various (modular) robotic components, as well as transfer of power and data. Finally, we provide a critical analysis of our work and outline its future work.

Abstract 5361: Deep learning models capture high dimensional features for cell morphology analysis

Abstract Many methods for imaging and sorting tumor cells require biomarker labels that perturb cells and may create a selection bias. The Deepcell platform can characterize and sort cells using only brightfield images of unlabeled single cells, thereby enabling more comprehensive and unbiased assessment of tumor cell morphology and heterogeneity. Cells are imaged with a high-speed camera during microfluidic flow and brightfield images are analyzed in real-time by deep learning models to generate quantitative embeddings, which are reproducible and multi-dimensional descriptions of cell morphology. One of the key technical challenges in building this platform was the development of an AI model to extract features from cell images from diverse human cells without prior knowledge of specific cell types, cell preparation, or other application-specific knowledge for an exploratory approach. Deepcell’s ‘Human Foundation Model’ (HFM) is a feature encoder that transforms cell images into 128 dimensional embedding vectors. The model backbone, responsible for extracting image features, is based on the ResNet18 convolutional deep neural network architecture. Training utilizes a multi-task semi-supervised training framework that combines the VicReg self-supervised learning model, which learns images features without labels, along with supervised auxiliary models using labeled data. These sub-models enable the backbone model to recognize specific cell attributes such as granulation, pigmentation, characteristics of malignancy, attributes related to cellular states like apoptosis and necrosis. We augment the deep learning embeddings with computer vision derived cellular features, such as area, perimeter, intensity, and texture features to improve model interpretability and accuracy. Here, we describe how the HFM self-supervised backbone model was trained, the discriminatory power added by the supervised tasks, and validation of the reproducibility and generalization capabilities of the resulting model. We also demonstrate how the resulting embeddings can be visualized using the Deepcell Cloud software and how clusters of cells in our embedding space can be related to cell images, interpretable morphological features, and actionable biological characteristics. Potential applications of the Deepcell platform are diverse and promising, such as label-free enrichment of malignant cells and discovery of morphologically distinct subpopulations of heterogeneous cancer cell samples. Citation Format: Kevin B. Jacobs, Kiran Saini, Nianzhen Li, Senzeyu Zhang, Ryan Carelli, Amy Wong-Thai, Stephane C. Boutet, Anastasia Mavropoulos, Andreja Jovic, Jeanette Mei, Mahyar Salek, Maddison Masaeli. Deep learning models capture high dimensional features for cell morphology analysis. [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 5361.

Academic program recommendations for graduate degrees in medical physics: AAPM Report No. 365 (Revision of Report No. 197).

Academic program recommendations for graduate degrees in medical physics: AAPM Report No. 365 (Revision of Report No. 197).

Model Assumptions and Data Characteristics: Impacts on Domain Adaptation in Building Segmentation

Studies on domain adaptation (DA) for remote sensing (RS) imagery analysis lack consistency in selection and description of evaluation scenarios. Without properly characterizing datasets, model assumptions, and evaluation scenarios, it is difficult to objectively compare DA methods and reach conclusions about their suitability across different applications. With this motivation, this work seeks to empirically assess to which extent the interaction between data characteristics and model assumptions influence the effectiveness of DA methods. Using the widely explored task of building footprint segmentation as case study, we perform a large-scale study across over 200 domain adaptation scenarios that include variations across view angles, areas observed, and sensors used for data acquisition. Rather than adopting different model architectures or optimization criteria, we contrast the performances of two DA methods based on adversarial learning that differ only in their assumptions about source and target domains. Informed by metadata and data characteristics unveiled using traditional computer vision techniques as well as pre-trained deep models, we provide a detailed meta-analysis of experiments highlighting the importance of accurately considering data assumptions for DA in RS segmentation tasks. As a “cherry-picking” exercise demonstrates, different claims regarding which model is best could be made by selecting different subsets of evaluation scenarios. While well-calibrated assumptions can be beneficial, mismatching assumptions can lead to negative biases in DA applications. This study intends to motivate the community towards more consistent evaluation protocols, while providing recommendations and insights toward creating novel benchmark datasets, documenting data characteristics, application-specific knowledge, and model assumptions.

A Systematic Methodology for Developing Bowtie in Risk Assessment: Application to Borescope Inspection

Background—Bowtie analysis is a broadly used tool in risk management to identify root causes and consequences of hazards and show barriers that can prevent or mitigate the events to happen. Limitations of the method are reliance on judgement and an ad hoc development process. Purpose—Systematic approaches are needed to identify threats and consequences, and to ascertain mitigation and prevention barriers. Results—A new conceptual framework is introduced by combining the Bowtie method with the 6M structure of Ishikawa to categorise the threats, consequences and barriers. The method is developed for visual inspection of gas turbine components, for which an example is provided. Originality—Provision of a more systematic methodology has the potential to result in more comprehensive Bowtie risk assessments, with less chance of serious omissions. The method is expected to find application in the broader industry, and to support operators who are non-risk experts but have application-specific knowledge, when performing Bowtie risk assessment.

Pangolin

There is growing interest in graph pattern mining (GPM) problems such as motif counting. GPM systems have been developed to provide unified interfaces for programming algorithms for these problems and for running them on parallel systems. However, existing systems may take hours to mine even simple patterns in moderate-sized graphs, which significantly limits their real-world usability. We present Pangolin , an efficient and flexible in-memory GPM framework targeting shared-memory CPUs and GPUs. Pangolin is the first GPM system that provides high-level abstractions for GPU processing. It provides a simple programming interface based on the extend-reduce-filter model, which allows users to specify application specific knowledge for search space pruning and isomorphism test elimination. We describe novel optimizations that exploit locality, reduce memory consumption, and mitigate the overheads of dynamic memory allocation and synchronization. Evaluation on a 28-core CPU demonstrates that Pangolin outperforms existing GPM frameworks Arabesque, RStream, and Fractal by 49×, 88×, and 80× on average, respectively. Acceleration on a V100 GPU further improves performance of Pangolin by 15× on average. Compared to state-of-the-art hand-optimized GPM applications, Pangolin provides competitive performance with less programming effort.

Rough set based lattice structure for knowledge representation in medical expert systems: Low back pain management case study

The aim of medical knowledge representation is to capture detailed domain knowledge in a clinically efficient manner, and to offer reliable resolution with the acquired knowledge. The knowledge base should allow incremental growth with inclusion of updated knowledge over the time. Accommodating the knowledge acquired from a variety of knowledge sources by different knowledge engineers may lead to a redundant and inconsistent design of the knowledge base, increasing the storage size and the time for knowledge retrieval. In this paper, we have proposed a rough set based lattice framework for representation of knowledge in medical expert systems which overcomes the problem of redundancy and inconsistency in knowledge and offers computational efficiency with respect to both time and space. The proposed knowledge representation offers a flexible scheme for expressing diverse possibilities of inter-relatedness among the symptoms and diseases in a systematic manner within the lattice structures.Through our design, we generate an optimal set of decision rules for use during inference. Reliability of each rule is measured using a new metric called credibility factor, and the certainty and coverage factors of a decision rule have been re-defined. During inference, the medical expert system considers the highly reliable and certain rules first, and the possible and uncertain rules at the later stages, if recommended by physicians. The proposed scheme ensures completeness, consistency, integrity, non-redundancy, and ease of access. These qualities are automatically preserved in the designed knowledge base while being updated with the new knowledge of medical advancements. As a result, the overall maintenance cost of a medical expert system is significantly reduced. The proposed knowledge representation technique has been illustrated using an example from the domain of low back pain.Though the proposed technique has been demonstrated for low back pain, it offers a generalized scheme for knowledge representation and may be adopted by medical expert systems (e.g. pathology, psychology, and many other specialties) having complexity similar to this one; no application-specific knowledge representation technique needs to be designed. The medical expert system developed using this scheme may act as a standardized example for development of different unexplored or semi-explored health care automated systems. This scheme may also be applied for pattern recognition, rule mining, and conflict analysis in complex medical data domains. Our design will have a wide ranging impact towards finding low-cost, reliable and available healthcare solutions for different diseases, especially in primary care units where expert physicians are scarce.

Determining Application-Specific Knowledge for Improving Robustness of Sequential Circuits

Due to their shrinking feature sizes as well as environmental influences, such as high-energy radiation, electrical noise, and particle strikes, integrated circuits are getting more vulnerable to transient faults. Accordingly, how to make those circuits more robust has become an essential step in today’s design flows. Methods increasing the robustness of circuits against these faults already exist for a long period of time but either introduce huge additional logic, change the timing behavior of the circuit, or are applicable for dedicated circuits such as microprocessors only. In this paper, we propose an alternative method, which overcomes these drawbacks by determining application-specific knowledge of the circuit, namely the relations of flip-flops and when they assume the same value. By this, we exploit partial redundancies, which are inherent in most circuits anyway (even the optimized ones), to frequently compare the circuit signals for their correctness—eventually leading to an increased robustness. Since determining the correspondingly needed information is a computationally hard task, formal methods, such as bounded model checking, satisfiability-based automatic test pattern generation, and binary decision diagrams, are utilized for this purpose. The resulting methodology requires only a slight increase in additional hardware, does only influence the timing behavior of the circuit negligibly, and is automatically applicable to arbitrary circuits. Experimental evaluations confirm these benefits.

Mind the gaps: Edge-based monitoring for continuous room-level location tracking

Abstract In-home monitoring applications often need strong and reliable contextual information to be of value. Indoor location tracking can play an important role in providing this context. However, most current indoor tracking solutions require extensive, application-specific knowledge and pre-deployment preparation, often including detailed information about floorplans and RF signal strength calibration, which is not practical in many deployments. This paper introduces an edge-based location monitoring technique that uses doorway sensors to detect doorway crossing events with walking directions. The approach is designed to provides room-level tracking. Users movement from a wearable device is used in conjunction with the edge-monitored approach to track multiple people. This method also enables the tracked location and doorway crossing events to be transformed into states and state transitions, respectively. This provides an opportunity to implement stochastic models to the edge-based location tracking to correct tracking errors and improve its accuracy, as demonstrated in simulation. This location tracking method has been implemented and deployed in a health monitoring study on dementia-related agitation.

An application-independent continuum mechanics interface for virtual engineering

Digitalization in industry applies to all product life cycle’s stages. An automation of the early conceptual design phase promises to make use of the maximum benefit, since decisions in the early development phases have the greatest potential for reducing the product’s life cycle costs (Pahl et al. in Engineering design, a systematic approach. Springer, London, 2007). There are several approaches to automate the early development phase: from more individual code or script-based solutions to formalized design methods such as knowledge-based engineering or model-based systems engineering. Running high-fidelity simulations leads to an improved decision basis in the early design phase, where the significant decisions, as the product architecture, are taken. This work shows a generic interface between the Unified Modeling Language or Java and continuum mechanics (CM) engineering tools. The interface introduces an application-independent layer containing all the (product) information to create, run and postprocess CM simulations. The boundary and physical conditions are modeled application independently in their generic representations. This approach offers the following advantages: first, an easy creation of engineering workflows through the automated generation and execution of engineering simulations. Second, the engineer is freed from the need for simulation application-specific knowledge. Finally, development costs are saved as commercial simulation tools, become arbitrarily interchangeable with open-source simulation tools.

An input-centric performance model for computational offloading of mobile applications

Computational offloading frameworks are a widely-researched technology for optimising mobile applications through the use of cloud resources. Existing frameworks fail to fully account for the effect of input data characteristics on application behaviour. Comprehensive timing models exist in the literature, but feature information requirements and performance overheads that preclude use on mobile devices. In this paper, we propose a conceptual model for an input-centric view of application performance. Our proposed model simplifies the existing count-and-weights and pipeline timing models to significantly reduce their information and processing requirements, facilitating use on resource-constrained mobile devices. Our proposed model also utilises symbolic execution techniques to account for the effects of application input data characteristics. Validation with both synthetic and real device datasets demonstrates that our model provides an extremely accurate approximation of the count-and-weights model. Results demonstrate the predictive power of our model for linear execution paths with no loops or recursion. Further work with improved symbolic execution techniques may look to expand application of our proposed model to real-world use cases. The proposed input-centric approach provides a promising foundation for incorporating a deeper level of application-specific knowledge into computational offloading framework cost models, with the potential to contribute to higher-quality offloading decisions.

Contain-ed

Network Functions Virtualization (NFV) has enabled operators to dynamically place and allocate resources for network services to match workload requirements. However, unbounded end-to-end (e2e) latency of Service Function Chains (SFCs) resulting from distributed Virtualized Network Function (VNF) deployments can severely degrade performance. In particular, SFC instantiations with inter-data center links can incur high e2e latencies and Service Level Agreement (SLA) violations. These latencies can trigger timeouts and protocol errors with latency-sensitive operations. Traditional solutions to reduce e2e latency involve physical deployment of service elements in close proximity. These solutions are, however, no longer viable in the NFV era. In this paper, we present our solution that bounds the e2e latency in SFCs and inter-VNF control message exchanges by creating micro-service aggregates based on the affinity between VNFs. Our system, Contain-ed, dynamically creates and manages affinity aggregates using light-weight virtualization technologies like containers, allowing them to be placed in close proximity and hence bounding the e2e latency. We have applied Contain-ed to the Clearwater IP Multimedia Subsystem and built a proof-of-concept. Our results demonstrate that, by utilizing application and protocol specific knowledge, affinity aggregates can effectively bound SFC delays and significantly reduce protocol errors and service disruptions.

Application-Guided Power-Efficient Fault Tolerance for H.264 Context Adaptive Variable Length Coding

This paper presents a fault-tolerance technique for H.264's Context-Adaptive Variable Length Coding (CAVLC) on unreliable computing hardware. The application-specific knowledge is leveraged at both algorithm and architecture levels to protect the CAVLC process (especially context adaptation and coding tables) in a reliable yet power-efficient manner. Specifically, the statistical analysis of coding syntax and video content properties are exploited for: (1) selective redundancy of coefficient/header data of video bitstreams; (2) partitioning the coding tables into various sub-tables to reduce the power overhead of fault tolerance; and (3) run-time power management of memory parts storing the sub-tables and their parity computations. Experimental results demonstrate that leveraging application-specific knowledge reduces area and performance overhead by 2x compared to a double-parity table protection technique. For functional verification and area comparison, the complete H.264 CAVLC architecture is prototyped on a Xilinx Virtex-5 FPGA (though not limited to it).

Diverging Library and Information Services, Converging Technologies, Surging Professional Roles and Emerging Scenario for LIS Manpower Development

Growing globalization, increasing competition and technological turbulences are demanding generation of new knowledge, more intensive search for ideas from all cultures and societies and adopt best practices. Institutions that generate new knowledge, accumulate the need and application specific knowledge and strategically use it are thriving and advancing in the present era. Libraries had always remained an important component of the knowledge infrastructure and are now required to undergo enormous transformation in view of the growing digital landscape, networked environment, and changes in ways users search, access and communicate information. Libraries have opportunities to expand their functionalities and offer services in newly diverging dimensions. While ICTs are converging for facilitating compact storage, quick retrieval and anytime and anywhere access to information from small handheld devices, a typical information user is facing information overload and time constrained environment and hence requires expert assistance in finding worthwhile information. The big data challenges, increasing unbundling of information, emerging new knowledge management (KM) tools, and content management software are offering new service opportunities for LIS professionals. How quickly professionals are able to identify service gaps, perceive and prepare for new roles, will determine the new service opportunities because when trained manpower is deployed at the right place and put to work, new possible services are automatically likely to take off. This paper presents the on-going changes in libraries, developments on information turf and their implications for information management. Discusses the surging roles and possible areas of service opportunities for LIS professionals. Portrays the opportunities for expansion and desired directions of change in the LIS education and training programs in India. In the wake of diverging functionalities of libraries and information centres, technological developments and surging professional roles, this study highlights the need for rejigging LIS education and training programs in India.

Learning-Based Compressive Subsampling

The problem of recovering a structured signal $\mathbf{x} \in \mathbb{C}^p$ from a set of dimensionality-reduced linear measurements $\mathbf{b} = \mathbf {A}\mathbf {x}$ arises in a variety of applications, such as medical imaging, spectroscopy, Fourier optics, and computerized tomography. Due to computational and storage complexity or physical constraints imposed by the problem, the measurement matrix $\mathbf{A} \in \mathbb{C}^{n \times p}$ is often of the form $\mathbf{A} = \mathbf{P}_{\Omega}\boldsymbol{\Psi}$ for some orthonormal basis matrix $\boldsymbol{\Psi}\in \mathbb{C}^{p \times p}$ and subsampling operator $\mathbf{P}_{\Omega}: \mathbb{C}^{p} \rightarrow \mathbb{C}^{n}$ that selects the rows indexed by $\Omega$. This raises the fundamental question of how best to choose the index set $\Omega$ in order to optimize the recovery performance. Previous approaches to addressing this question rely on non-uniform \emph{random} subsampling using application-specific knowledge of the structure of $\mathbf{x}$. In this paper, we instead take a principled learning-based approach in which a \emph{fixed} index set is chosen based on a set of training signals $\mathbf{x}_1,\dotsc,\mathbf{x}_m$. We formulate combinatorial optimization problems seeking to maximize the energy captured in these signals in an average-case or worst-case sense, and we show that these can be efficiently solved either exactly or approximately via the identification of modularity and submodularity structures. We provide both deterministic and statistical theoretical guarantees showing how the resulting measurement matrices perform on signals differing from the training signals, and we provide numerical examples showing our approach to be effective on a variety of data sets.

Automated walking aid detector based on indoor video recordings.

Due to the rapidly aging population, developing automated home care systems is a very important step in taking care of elderly people. This will enable us to automatically monitor the health of senior citizens in their own living environment and prevent problems before they happen. One of the challenging tasks is to actively monitor walking habits of elderlies, who alternate between the use of different walking aids, and to combine this with automated fall risk assessment systems. We propose a camera based system that uses object categorization techniques to robustly detect walking aids, like a walker, in order to improve the classification of the fall risk. By automatically integrating the application specific scenery knowledge like camera position and used walker type, we succeed in detecting walking aids within a single frame with an accuracy of 68% for trajectory A and 38% for trajectory B. Furthermore, compared to current state of the art detection systems, we use a rather limited set of training data to achieve this accuracy and thus create a system that is easily adaptable for other applications. Moreover, we applied spatial constraints between detections to optimize the object detection output and to limit the amount of false positive detections. Finally, we evaluate the output on a walking sequence base, leading up to a 92.3% correct classification rate of walking sequences. It can be noted that adapting this approach to other walking aids, like a walking cane, is quite straightforward and opens up the door for many future applications.

The Baquara2 knowledge-based framework for semantic enrichment and analysis of movement data

The analysis of movements frequently requires more than just spatio-temporal data. Thus, despite recent progresses in trajectory handling, there is still a gap between movement data and formal semantics. This gap hinders movement analyses benefiting from available knowledge, with well-defined and widely agreed semantics. This article describes the Baquara2 framework to help narrow this gap by exploiting knowledge bases to semantically enrich and analyze movement data. It provides an ontological model for structuring and abstracting movement data in a multilevel hierarchy of progressively detailed movement segments that generalize concepts such as trajectories, stops, and moves. Baquara2 also includes a general customizable process to annotate movement data with concepts and objects described in ontologies and Linked Open Data (LOD) collections. The resulting semantic annotations enable queries for movement analyses based on application and domain specific knowledge. The proposed framework has been used in experiments to semantically enrich movement data collected from social media with geo-referenced LOD. The obtained results enable powerful queries that illustrate Baquara2 capabilities.

Extracting meaning from biological imaging data.

Biological imaging continues to improve, capturing continually longer-term, richer, and more complex data, penetrating deeper into live tissue. How do we gain insight into the dynamic processes of disease and development from terabytes of multidimensional image data? Here I describe a collaborative approach to extracting meaning from biological imaging data. The collaboration consists of teams of biologists and engineers working together. Custom computational tools are built to best exploit application-specific knowledge in order to visualize and analyze large and complex data sets. The image data are summarized, extracting and modeling the features that capture the objects and relationships in the data. The summarization is validated, the results visualized, and errors corrected as needed. Finally, the customized analysis and visualization tools together with the image data and the summarization results are shared. This Perspective provides a brief guide to the mathematical ideas that rigorously quantify the notion of extracting meaning from biological image, and to the practical approaches that have been used to apply these ideas to a wide range of applications in cell and tissue optical imaging.