Human-Centered Computing at Khoury College of Computer Sciences

Supporting Modular Design of Machine-Knitting Programs

Knitting machines can manufacture complex layered, textured, and multi-material fabrics and garments. With new programming languages and interfaces there is greater access to the machine’s capabilities. Developers can now create machine instructions that produce a fabric sample or garment. Such files can be easily shared with others, but modifying and combining these samples requires extensive expertise in knitting-specific programming languages, substantial effort, and time. Knit programming offers little support for modular design. We take a step towards modular knitting-machine programming and present QUILT: Quality Unification Infrastructure for Loop-based Textiles. QUILT enables knit programmers to create swatches from knitting programs and lay these swatches out spatially on a 2-dimensional grid. We use three novel knit-program merging algorithms to merge the connected swatches into a quilt program. The knitted structures of each swatch remain unchanged, and our algorithms ensure that the swatches are joined by a seamless boundary that maintains the constraints of knitting-machine programming and knitted-structure construction.

Cognitive Screening Through Brand Recognition

This project explores an innovative approach to cognitive health screening using familiar brand logos as cues for memory and language. We have developed NEURO-logo (Neurological Screening via Logo Recognition), a tool that assesses naming and recognition of widely known logos to detect early signs of cognitive or linguistic impairment. Unlike tests that rely on faces, written words, or generic objects, logos offer a unique advantage in that they are culturally salient, visually distinctive, and semantically rich, thus engaging multiple cognitive domains such as visual recognition, semantic memory, language retrieval, and brand association. Because logos are encountered repeatedly in daily life, they may offer a more ecologically valid and emotionally neutral stimulus set for screening subtle cognitive decline.

We will initially gather normative data in healthy adults to establish baseline performance followed by testing in individuals with known neurological conditions to assess diagnostic sensitivity. NEURO-logo aims to provide an engaging, low-burden, and scalable screening method that brings cognitive assessment closer to everyday life.

Accessibility in ability-diverse collaboration

Khoury researchers are investigating how accessibility is created and negotiated within ability-diverse teams in the contexts of collaborative writing, creative making, ideation, and remote work. Some examples of our recent projects include developing auditory techniques for enhancing accessibility in asynchronous and synchronous collaborative writing, an audio-enhanced loom for accessible weaving, and an audio-tactile system for accessible pattern generation. Through our work, we also critically interrogate what roles technologies play in reshaping group dynamics and redistributing the labor of creating access in ability-diverse teams. (PI: Maitraye Das)

The Physical Activity Using Wearable Sensors (PAAWS) study

Accurate measurement of human behavior using devices could significantly advance current knowledge on the dose-response relationships between chronic diseases and behaviors such as physical activity, sedentary behavior, and sleep. The objective of this project is to develop new algorithms to accurately measure behavior 24/7 using wearable sensors. We are collecting wearable sensor datasets from people as they go about their normal lives, labeling the data second-by-second, and using the labeled data to develop and test new machine learning algorithms that will more reliably detect activities and habits. We aim to help the larger research community perform comparisons between algorithms on realistic datasets of behavior. (PI: Stephen Intille)