Following the Clues: Experiments on Person Re-ID using Cross-Modal Intelligence
This addresses privacy concerns for pedestrians in autonomous driving and AI research by incrementally improving person re-identification with interpretable PII detection.
The paper tackles the problem of privacy risks from personally identifiable information (PII) in street-level datasets by proposing cRID, a cross-modal framework that detects textual describable clues to enhance person re-identification, showing improved performance in cross-dataset scenarios such as from Market-1501 to CUHK03-np.
The collection and release of street-level recordings as Open Data play a vital role in advancing autonomous driving systems and AI research. However, these datasets pose significant privacy risks, particularly for pedestrians, due to the presence of Personally Identifiable Information (PII) that extends beyond biometric traits such as faces. In this paper, we present cRID, a novel cross-modal framework combining Large Vision-Language Models, Graph Attention Networks, and representation learning to detect textual describable clues of PII and enhance person re-identification (Re-ID). Our approach focuses on identifying and leveraging interpretable features, enabling the detection of semantically meaningful PII beyond low-level appearance cues. We conduct a systematic evaluation of PII presence in person image datasets. Our experiments show improved performance in practical cross-dataset Re-ID scenarios, notably from Market-1501 to CUHK03-np (detected), highlighting the framework's practical utility. Code is available at https://github.com/RAufschlaeger/cRID.