Unified Representation Learning for Cross Model Compatibility
This addresses the practical problem of enabling visual search systems to work across different embedding models without re-encoding user images, which is important for privacy-preserving applications.
The paper tackles the Cross Model Compatibility problem in visual search applications by proposing a unified representation learning framework with a Residual Bottleneck Transformation module and new training scheme, achieving significant performance improvements over previous approaches in face recognition and person re-identification scenarios.
We propose a unified representation learning framework to address the Cross Model Compatibility (CMC) problem in the context of visual search applications. Cross compatibility between different embedding models enables the visual search systems to correctly recognize and retrieve identities without re-encoding user images, which are usually not available due to privacy concerns. While there are existing approaches to address CMC in face identification, they fail to work in a more challenging setting where the distributions of embedding models shift drastically. The proposed solution improves CMC performance by introducing a light-weight Residual Bottleneck Transformation (RBT) module and a new training scheme to optimize the embedding spaces. Extensive experiments demonstrate that our proposed solution outperforms previous approaches by a large margin for various challenging visual search scenarios of face recognition and person re-identification.