Computational Methodologies and Methods for Visual Research

Computational tools, machine learning, and computer vision are rapidly entering the visual researcher's toolkit. These methods can analyse images at scales and speeds previously unimaginable, surfacing patterns across vast and varied visual collections. At the same time, newer models open possibilities for interpretive and qualitative work, assisting with annotation, coding, comparison, and the close reading of images. This session asks how such tools can be integrated into visual research without surrendering the field's critical, reflexive, and theoretically grounded commitments.

We invite contributions that develop, apply, or critically interrogate computational methodologies and methods for visual research, whether as broad methodological approaches or as more specialised techniques, across the qualitative and quantitative spectrum. These might include large-scale image classification and clustering, automated content analysis, machine learning for interpretive coding and hermeneutic analysis, or mixed-methods designs that combine computational scale with situated reading. We also welcome reflections on validity, reproducibility, and inference in machine-assisted visual work. Equally welcome are papers that examine the assumptions built into these methods, including how models work as interpretive instruments, what their outputs obscure or distort, the influence of training data, and the questions of bias and research ethics they raise.

This session works toward a critical methodological agenda for visual research where both images and the means of analysing them are increasingly computational. We welcome contributions from across disciplines, including data science, computational social science, and the digital humanities, and especially work that connects methodological innovation to substantive research questions.