Lightweight Attribute Localizing Models for Pedestrian Attribute Recognition

Pedestrian attribute recognition (PAR) focuses on identifying attributes in pedestrian images, with applications in person retrieval, suspect reidentification, and soft biometrics. However, neural networks for PAR suffer from overparameterization and high computational complexity, making them unsuitable for resource_constrained devices. Tensor_based compression methods factorize layers without preserving the gradient direction during compression, leading to inefficient compression and an accuracy loss. We propose a novel approach for determining the optimal ranks of low_rank layers, ensuring that the gradient direction of the compressed model aligns with that of the original model. This means that the compressed model 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 attribute localization model, followed by compression using canonical polyadic decomposition with error_preserving correction or singular value decomposition. This results in a reduction in model complexity while maintaining high performance.

Authors

• Ashish Jha
• Dmitrii Ermilov
• Anh Huy Phan
• Konstantin Sobolev FusionBrain Lab Lomonosov Moscow State University
• Salman Ahmadi-Asl
• Naveed Ahmed
• Imran Junejo
• Zaher Al Aghbari
• Thar Baker
• Ahmed M. Khedr
• Andrzej Cichocki

Keywords

• Tensors
• Convolutional neural networks
• Computational modeling
• Image coding
• Location awareness
• Image recognition
• Human activity recognition
• Attribute Recognition
• Pedestrian Attribute
• Pedestrian Attribute Recognition
• Convolutional Neural Network
• Object Detection
• Singular Value Decomposition
• Light Weight
• Accuracy Loss
• Gradient Direction
• Direct Model
• Pose Estimation
• Compression Method
• Surveillance Cameras
• Binary Search
• Low-rank Approximation
• Tensor Decomposition
• Compression Efficiency
• Reduce Model Complexity
• Low-rank Tensor
• Truncated Singular Value Decomposition
• Tucker Decomposition
• Convolutional Layers
• Norm Minimization
• Convolution Kernel
• Original Network
• Image X
• Tensor Rank
• Holistic Method
• F1 Score
• Component Loadings