Lightweight Attribute Localizing Models for Pedestrian Attribute Recognition
This work addresses the need for lightweight models for Pedestrian Attribute Recognition on resource-constrained devices, representing an incremental improvement in compression methods.
The paper tackles the problem of over-parameterization and high computational complexity in Deep Neural Networks for Pedestrian Attribute Recognition by proposing a novel approach to determine optimal ranks for low-rank layers, ensuring gradient direction preservation during compression, which reduces model complexity while maintaining high performance.
Pedestrian Attribute Recognition (PAR) focuses on identifying various attributes in pedestrian images, with key applications in person retrieval, suspect re-identification, and soft biometrics. However, Deep Neural Networks (DNNs) for PAR often suffer from over-parameterization and high computational complexity, making them unsuitable for resource-constrained devices. Traditional tensor-based compression methods typically factorize layers without adequately preserving the gradient direction during compression, leading to inefficient compression and a significant accuracy loss. In this work, we propose a novel approach for determining the optimal ranks of low-rank layers, ensuring that the gradient direction of the compressed model closely aligns with that of the original model. This means that the compressed model effectively preserves the update direction of the full model, enabling more efficient compression for PAR tasks. The proposed procedure optimizes the compression ranks for each layer within the ALM model, followed by compression using CPD-EPC or truncated SVD. This results in a reduction in model complexity while maintaining high performance.