We introduce multiXview, an interactive retrieval framework for synchronized multi-camera video collections. It features a multi-index search engine that supports natural-language queries over visual embeddings, speech transcripts, and scene descriptions. It supports a synchronized multi-stream player offering parallel playback, and a timeline-based navigation view for temporal scoping and faceted exploration. These components address the redundancy and fragmentation of overlapping egocentric and exocentric video feeds and enable users to locate, aggregate, and reconstruct events across partial perspectives. This paper focuses on system design and implementation, with quantitative and qualitative evaluation to take place at the CASTLE 2025 Grand Challenge Interactive Track.

Abstract Lifelogging is a form of personal data collection which seeks to capture the totality of one’s experience through intelligent technology and sensors. Yet despite notable advancement in such technologies, there remain persistent challenges to developing interactive systems to analyse the types of large-scale personal collections often generated by lifelogging. In response to this, we present the Lifelog Application Design (LAD) model which is intended to address these challenges and support the design of more novel interactive systems that may target a broader range of application use cases. The model is deliberately structured to remain impartial to the specific personal data, technology platform, or application criterion, to provide maximum utility across the domain. We demonstrate this utility by exploring two case studies and a retrospective analysis of VRLE, a real-world application prototype developed to examine the potential of large-scale personal data retrieval in virtual reality. This work is based on the accumulation of insights garnered from involvement in a number of collaborative lifelogging projects over the past decade. It is our goal to encourage future researchers to utilise the LAD model to support the design and development of their own application prototypes and further solidify the model’s contribution to the domain as a whole.

This paper describes the performance optimisation of a state-of-the-art relational database to more efficiently serve data for multimedia visualisations in the ViRMA prototype. We describe the baseline database and queries, along with two major optimisation steps that improve query efficiency, at the cost of slowing down dynamic updates. We evaluate the optimisations with a case study of a lifelog collection of 182K images, showing that the time to produce complex visualisations is reduced by orders of magnitude.

In this paper we describe the latest iteration of the Virtual Reality Multimedia Analytics (ViRMA) system, a novel approach to multimedia analysis in virtual reality which is supported by the Multi-dimensional Multimedia Model.