Visual Chronicles: Using Multimodal LLMs to Analyze Massive Collections of Images
It addresses the challenge of open-ended visual analysis for large-scale temporal image datasets, which is incremental as it adapts MLLMs to a new application.
The paper tackles the problem of analyzing massive collections of images over time to discover patterns, such as frequent co-occurring changes in a city, using Multimodal LLMs (MLLMs) without predetermined targets or labels, and finds that the system significantly outperforms baselines and identifies trends like 'addition of outdoor dining'.
We present a system using Multimodal LLMs (MLLMs) to analyze a large database with tens of millions of images captured at different times, with the aim of discovering patterns in temporal changes. Specifically, we aim to capture frequent co-occurring changes ("trends") across a city over a certain period. Unlike previous visual analyses, our analysis answers open-ended queries (e.g., "what are the frequent types of changes in the city?") without any predetermined target subjects or training labels. These properties cast prior learning-based or unsupervised visual analysis tools unsuitable. We identify MLLMs as a novel tool for their open-ended semantic understanding capabilities. Yet, our datasets are four orders of magnitude too large for an MLLM to ingest as context. So we introduce a bottom-up procedure that decomposes the massive visual analysis problem into more tractable sub-problems. We carefully design MLLM-based solutions to each sub-problem. During experiments and ablation studies with our system, we find it significantly outperforms baselines and is able to discover interesting trends from images captured in large cities (e.g., "addition of outdoor dining,", "overpass was painted blue," etc.). See more results and interactive demos at https://boyangdeng.com/visual-chronicles.