CPU-Oriented Multi-Cue Person Re-ID Soft-Biometric Fusion

This dataset contains person images captured by non-overlapping surveillance cameras for research on person re-identification. The dataset is designed for evaluating person re-identification methods in realistic multi-camera environments where individuals appear across different viewpoints, illumination conditions, poses, and background contexts, without camera overlap. It is especially intended for studying CPU-efficient person re-identification, soft-biometric fusion, and distributed model updating strategies under constrained computational settings. The image set includes samples of the same individuals observed from different non-overlapping cameras, enabling the analysis of cross-camera appearance variation and identity matching performance. This dataset may be used for benchmarking re-identification approaches, validating lightweight models, and supporting reproducible research in intelligent video surveillance and computer vision.