Multi Positive Contrastive Learning with Pose-Consistent Generated Images
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910398377099264 |
|---|---|
| author | Inayoshi, Sho Widya, Aji Resindra Ozaki, Satoshi Otsuka, Junji Ohashi, Takeshi |
| author_facet | Inayoshi, Sho Widya, Aji Resindra Ozaki, Satoshi Otsuka, Junji Ohashi, Takeshi |
| contents | Model pre-training has become essential in various recognition tasks. Meanwhile, with the remarkable advancements in image generation models, pre-training methods utilizing generated images have also emerged given their ability to produce unlimited training data. However, while existing methods utilizing generated images excel in classification, they fall short in more practical tasks, such as human pose estimation. In this paper, we have experimentally demonstrated it and propose the generation of visually distinct images with identical human poses. We then propose a novel multi-positive contrastive learning, which optimally utilize the previously generated images to learn structural features of the human body. We term the entire learning pipeline as GenPoCCL. Despite using only less than 1% amount of data compared to current state-of-the-art method, GenPoCCL captures structural features of the human body more effectively, surpassing existing methods in a variety of human-centric perception tasks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_03256 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Multi Positive Contrastive Learning with Pose-Consistent Generated Images Inayoshi, Sho Widya, Aji Resindra Ozaki, Satoshi Otsuka, Junji Ohashi, Takeshi Computer Vision and Pattern Recognition Model pre-training has become essential in various recognition tasks. Meanwhile, with the remarkable advancements in image generation models, pre-training methods utilizing generated images have also emerged given their ability to produce unlimited training data. However, while existing methods utilizing generated images excel in classification, they fall short in more practical tasks, such as human pose estimation. In this paper, we have experimentally demonstrated it and propose the generation of visually distinct images with identical human poses. We then propose a novel multi-positive contrastive learning, which optimally utilize the previously generated images to learn structural features of the human body. We term the entire learning pipeline as GenPoCCL. Despite using only less than 1% amount of data compared to current state-of-the-art method, GenPoCCL captures structural features of the human body more effectively, surpassing existing methods in a variety of human-centric perception tasks. |
| title | Multi Positive Contrastive Learning with Pose-Consistent Generated Images |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2404.03256 |