davanstrien/sam3-segment-test

Image Segmentation: Photograph using SAM3

This dataset contains semantic segmentation maps for photograph segmented in images from davanstrien/newspapers-with-images-after-photography using Meta's SAM3.

Generated using: uv-scripts/sam3 segmentation script

Statistics

• Objects Segmented: photograph
• Total Instances: 0
• Images with Detections: 0 / 10 (0.0%)
• Average Instances per Image: 0.00
• Output Format: semantic-mask

Processing Details

• Source Dataset: davanstrien/newspapers-with-images-after-photography
• Model: facebook/sam3
• Processing Date: 2026-03-30 15:54 UTC
• Processing Time: 0.3 minutes
• Confidence Threshold: 0.3
• Mask Threshold: 0.5
• Includes Bounding Boxes: Yes

Reproduction

uv run https://huggingface.co/datasets/uv-scripts/sam3/raw/main/segment-objects.py \
    davanstrien/newspapers-with-images-after-photography \
    <output-dataset> \
    --class-name photograph \
    --output-format semantic-mask \
    --confidence-threshold 0.3 \
    --mask-threshold 0.5  --include-boxes

---

Generated with UV Scripts