ser_balanced / README.md
marco-willi's picture
Upload README.md with huggingface_hub
58c10d3 verified
metadata
license: cdla-permissive-1.0
task_categories:
  - image-classification
tags:
  - camera-trap
  - wildlife
  - serengeti
  - snapshot-safari
  - megadetector
pretty_name: Snapshot Safari SER Balanced  Classroom Subset v1.0
size_categories:
  - 1K<n<10K

Snapshot Safari SER (Serengeti) — Balanced Classroom Subset v1.0

Summary

A curated, balanced subset of the Snapshot Safari 2024 Expansion SER (Serengeti National Park) camera trap dataset, prepared for use in the CAS Deep Learning — Computer Vision course exercises.

Archive Images Description
ser_balanced.tar.gz 1,850 Balanced, ≤200/class, full frames

Source

MegaDetector

Pre-computed MegaDetector v1000-redwood RDE-filtered results from LILA Science: snapshot-safari-2024-expansion-SER-subset-v1000.0.0-redwood_detections.threshold.filtered.json.zip

Used to filter frames (conf ≥ 0.8) and select the best frame per sequence.

Species

buffalo, elephant, empty, gazellegrants, gazellethomsons, hartebeest, impala, warthog, wildebeestblue, zebraplains

Statistics

Class Train Val Test Total
buffalo 140 30 30 200
elephant 140 30 30 200
empty 35 8 7 50
gazellegrants 140 30 30 200
gazellethomsons 140 30 30 200
hartebeest 140 30 30 200
impala 140 30 30 200
warthog 140 30 30 200
wildebeestblue 140 30 30 200
zebraplains 140 30 30 200
Total 1295 278 277 1850

Note: ~56% of images are IR/night (near-infrared, nearly greyscale).

Curation Details

  • Deduplication: one image per sequence (highest MD animal confidence frame)
  • Animal filter: MD animal confidence ≥ 0.8
  • Empty filter: max MD animal confidence < 0.2
  • Split strategy: stratified 70/15/15 by sequence ID — no sequence spans splits
  • Image resolution: resized to max 1024 px on longer side, JPEG quality 92
  • Format: ImageFolder layout — <split>/<label>/<filename>.jpg

Usage

from huggingface_hub import hf_hub_download
import tarfile

archive = hf_hub_download(
    "marco-willi/ser_balanced",
    "ser_balanced.tar.gz",
    repo_type="dataset",
)
with tarfile.open(archive) as tar:
    tar.extractall(DATA_PATH)
# → DATA_PATH/ser/ser_balanced/{train,val,test}/<label>/*.jpg