Learning Disentangled Attribute Representations for Robust Pedestrian Attribute Recognition

Abstract

Although various methods have been proposed for pedestrian attribute recognition, most studies follow the same feature learning mechanism, \ie, learning a shared pedestrian image feature to classify multiple attributes. However, this mechanism leads to low-confidence predictions and non-robustness of the model in the inference stage. In this paper, we investigate why this is the case. We mathematically discover that the central cause is that the optimal shared feature cannot maintain high similarities with multiple classifiers simultaneously in the context of minimizing classification loss. In addition, this feature learning mechanism ignores the spatial and semantic distinctions between different attributes. To address these limitations, we propose a novel disentangled attribute feature learning (DAFL) framework to learn a disentangled feature for each attribute, which exploits the semantic and spatial characteristics of attributes. The framework mainly consists of learnable semantic queries, a cascaded semantic-spatial cross-attention (SSCA) module, and a group attention merging (GAM) module. Specifically, based on learnable semantic queries, the cascaded SSCA module iteratively enhances the spatial localization of attribute-related regions and aggregates region features into multiple disentangled attribute features, used for classification and updating learnable semantic queries. The GAM module splits attributes into groups based on spatial distribution and utilizes reliable group attention to supervise query attention maps. Experiments on PETA, RAPv1, PA100k, and RAPv2 show that the proposed method performs favorably against state-of-the-art methods.

UniHCP: A Unified Model for Human-Centric Perceptions

“…With finetuning, our UniHCP achieves new SOTAs on nine out of the total twelve datasets and on par performance on the rest three datasets, even without task-specific design in architecture or task-specific priors, showing that Uni-HCP extracts complementary knowledge among humancentric tasks. Concretely, Table 4 shows that in the human attribute recognition task, UniHCP significantly surpasses previous SOTA DAFL [29] by +3.79% mA on PA-100K and +1.20% mA on RAPv2 datasets, respectively, which indicates that UniHCP well extracts the shared attribute information among using the output unit of global probabilities in the interpreter. Second, UniHCP also pushes the performance of another important human task, i.e., human parsing, to a new level.…”

“…The whole training takes 120 hours in total on 88 NVIDIA V100 GPUs. Method PA-100K RAPv2 SSC [28] 81.87 -C-Tran [36] 81.53 -Q2L [60] 80.72 -L2L [46] 82.37 -DAFL [29] 83…”

Correlation Graph Convolutional Network for Pedestrian Attribute Recognition

“…To show the effectiveness of the proposed network, we compare it with several state-of-the-art PAR methods: HPNet [55], VeSPA [59], JRL [60], PGDM [30], MsVAA [37], GRL [31], VAC [15], JLPSL-PAA [61], RA [62], ALM [23], MT-CAS [38], DTM+AWK [63], JLAC [39], Baseline [32], SSC [64], DAFL [65], and VTB [47].…”

Attribute correlation mask fusion network for pedestrian attribute recognition

“…To demonstrate the effectiveness of the proposed network, we compared it with several competing PAR methods, including: HPNet [33], VeSPA [34], PGDM [5], MsVAA [34], GRL [6], JLPLS-PAA [35], ALM [14], MT-CAS [10], DTM+AWK [36], JLAC [37], Baseline [29], SSC [28], Label2Label [38], DAFL [39], VTB [40] and EAM [12].…”