Detect-Person / README.md
meaculpitt's picture
init: placeholder README, training in progress
373bec2 verified
metadata
tags:
  - element_type:detect
  - model:yolov11-nano
  - object:person
manako:
  description: >-
    YOLOv11-nano for ground-level CCTV person detection on SN44 (training in
    progress)
  source: meaculpitt/Detect-Person
  prompt_hints: null
  input_payload:
    - name: frame
      type: image
      description: RGB frame
  output_payload:
    - name: detections
      type: detections
      description: List of bounding boxes for detected persons
  evaluation_score: null
  last_benchmark: null

Detect-Person — SN44

YOLOv11-nano fine-tuned for ground-level CCTV person detection.

Status: Training in progress. Model weights will be uploaded on completion.

  • Training data: CrowdHuman (15k) + BDD100K pedestrians (3.2k)
  • Validation: COCO val2017 persons
  • Input size: 1280×1280
  • Conf threshold: 0.35 (precision-tuned for false_positive pillar)