PhotoGuider releases new method
The story so far
PhotoGuider has introduced a large language model method that combines composition-aware cropping with 360° perspective selection. This method uses shared composition-aware tokens to select the viewing direction and frame 360° panoramic images, while also outputting camera parameters with explanatory scores. This method achieved the highest average mIoU performance on two standard 2D cropping benchmarks and, in a blinded ranking study of 100 panoramic scenes, its framing results were superior to the other two methods in 53.2% of cases.
AI-written from the reports · Updated 15 days ago
Coverage timeline
Follow the coverage to see every side of the story.
- Hugging Face Daily PapersPhotoGuider: Composition-Aware Cropping and 360^{circ} View Framing with Explainable Photographic Assessments
PhotoGuider uses shared composition perception tokens to select the viewing direction and frame 360° panoramic images, while outputting camera parameters with explanatory scores. This method achieves the highest average mIoU performance on two standard 2D cropping benchmarks, and in a blind ranking study evaluating 100 panoramic scenes, its framing results were better than the other two methods in 53.2% of cases.