A Quantitative Evaluation of the Expressivity of BMI, Pose and Gender in Body Embeddings for Recognition and Identification
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866908353429504000 |
|---|---|
| author | Pal, Basudha Huang, Siyuan Chellappa, Rama |
| author_facet | Pal, Basudha Huang, Siyuan Chellappa, Rama |
| contents | Person Re-identification (ReID) systems that match individuals across images or video frames are essential in many real-world applications. However, existing methods are often influenced by attributes such as gender, pose, and body mass index (BMI), which vary in unconstrained settings and raise concerns related to fairness and generalization. To address this, we extend the notion of expressivity, defined as the mutual information between learned features and specific attributes, using a secondary neural network to quantify how strongly attributes are encoded. Applying this framework to three ReID models, we find that BMI consistently shows the highest expressivity in the final layers, indicating its dominant role in recognition. In the last attention layer, attributes are ranked as BMI > Pitch > Gender > Yaw, revealing their relative influences in representation learning. Expressivity values also evolve across layers and training epochs, reflecting a dynamic encoding of attributes. These findings demonstrate the central role of body attributes in ReID and establish a principled approach for uncovering attribute driven correlations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_06451 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | A Quantitative Evaluation of the Expressivity of BMI, Pose and Gender in Body Embeddings for Recognition and Identification Pal, Basudha Huang, Siyuan Chellappa, Rama Computer Vision and Pattern Recognition Person Re-identification (ReID) systems that match individuals across images or video frames are essential in many real-world applications. However, existing methods are often influenced by attributes such as gender, pose, and body mass index (BMI), which vary in unconstrained settings and raise concerns related to fairness and generalization. To address this, we extend the notion of expressivity, defined as the mutual information between learned features and specific attributes, using a secondary neural network to quantify how strongly attributes are encoded. Applying this framework to three ReID models, we find that BMI consistently shows the highest expressivity in the final layers, indicating its dominant role in recognition. In the last attention layer, attributes are ranked as BMI > Pitch > Gender > Yaw, revealing their relative influences in representation learning. Expressivity values also evolve across layers and training epochs, reflecting a dynamic encoding of attributes. These findings demonstrate the central role of body attributes in ReID and establish a principled approach for uncovering attribute driven correlations. |
| title | A Quantitative Evaluation of the Expressivity of BMI, Pose and Gender in Body Embeddings for Recognition and Identification |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.06451 |