PhotoBench: Beyond Visual Matching Towards Personalized Intent-Driven Photo Retrieval

TL;DR

PhotoBench introduces a new benchmark for personalized photo retrieval that emphasizes multi-source reasoning over visual matching, highlighting the need for advanced agentic systems to handle complex, intent-driven queries in personal albums.

Contribution

It presents the first authentic, personal album-based benchmark for intent-driven photo retrieval, emphasizing multi-source reasoning and exposing limitations of current unified embedding models.

Findings

Unified embedding models struggle with non-visual constraints.

Agentic systems perform poorly in tool orchestration.

Next frontier requires robust multi-source reasoning systems.

Abstract

Personal photo albums are not merely collections of static images but living, ecological archives defined by temporal continuity, social entanglement, and rich metadata, which makes the personalized photo retrieval non-trivial. However, existing retrieval benchmarks rely heavily on context-isolated web snapshots, failing to capture the multi-source reasoning required to resolve authentic, intent-driven user queries. To bridge this gap, we introduce PhotoBench, the first benchmark constructed from authentic, personal albums. It is designed to shift the paradigm from visual matching to personalized multi-source intent-driven reasoning. Based on a rigorous multi-source profiling framework, which integrates visual semantics, spatial-temporal metadata, social identity, and temporal events for each image, we synthesize complex intent-driven queries rooted in users' life trajectories. Extensive evaluation on PhotoBench exposes two critical limitations: the modality gap, where unified embedding models collapse on non-visual constraints, and the source fusion paradox, where agentic systems perform poor tool orchestration. These findings indicate that the next frontier in personal multimodal retrieval lies beyond unified embeddings, necessitating robust agentic reasoning systems capable of precise constraint satisfaction and multi-source fusion. Our PhotoBench is available.