Aligning Object Detector Bounding Boxes with Human Preference
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_ | 1866911996409020416 |
|---|---|
| author | Strafforello, Ombretta Kayhan, Osman S. Inel, Oana Schutte, Klamer van Gemert, Jan |
| author_facet | Strafforello, Ombretta Kayhan, Osman S. Inel, Oana Schutte, Klamer van Gemert, Jan |
| contents | Previous work shows that humans tend to prefer large bounding boxes over small bounding boxes with the same IoU. However, we show here that commonly used object detectors predict large and small boxes equally often. In this work, we investigate how to align automatically detected object boxes with human preference and study whether this improves human quality perception. We evaluate the performance of three commonly used object detectors through a user study (N = 123). We find that humans prefer object detections that are upscaled with factors of 1.5 or 2, even if the corresponding AP is close to 0. Motivated by this result, we propose an asymmetric bounding box regression loss that encourages large over small predicted bounding boxes. Our evaluation study shows that object detectors fine-tuned with the asymmetric loss are better aligned with human preference and are preferred over fixed scaling factors. A qualitative evaluation shows that human preference might be influenced by some object characteristics, like object shape. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2408_10844 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Aligning Object Detector Bounding Boxes with Human Preference Strafforello, Ombretta Kayhan, Osman S. Inel, Oana Schutte, Klamer van Gemert, Jan Computer Vision and Pattern Recognition Previous work shows that humans tend to prefer large bounding boxes over small bounding boxes with the same IoU. However, we show here that commonly used object detectors predict large and small boxes equally often. In this work, we investigate how to align automatically detected object boxes with human preference and study whether this improves human quality perception. We evaluate the performance of three commonly used object detectors through a user study (N = 123). We find that humans prefer object detections that are upscaled with factors of 1.5 or 2, even if the corresponding AP is close to 0. Motivated by this result, we propose an asymmetric bounding box regression loss that encourages large over small predicted bounding boxes. Our evaluation study shows that object detectors fine-tuned with the asymmetric loss are better aligned with human preference and are preferred over fixed scaling factors. A qualitative evaluation shows that human preference might be influenced by some object characteristics, like object shape. |
| title | Aligning Object Detector Bounding Boxes with Human Preference |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2408.10844 |