Use Case and High-Level Description

This is a pedestrian detector based on backbone with hyper-feature + R-FCN for the Retail scenario.




Metric Value
AP 80.14%
Pose coverage Standing upright, parallel to image plane
Support of occluded pedestrians YES
Occlusion coverage <50%
Min pedestrian height 80 pixels (on 1080p)
Max objects to detect 200
GFlops 12.427
MParams 3.244
Source framework Caffe*

Average Precision (AP) is defined as an area under the precision/recall curve. Validation dataset consists of ~50K of images from ~100 different scenes.



  1. name: data , shape: [1x3x544x992] - An input image in following format [1xCxHxW]. The expected channel order is BGR.
  2. name: im_info, shape: [1x6] - An image information [544, 992, 992/frame_width, 544/frame_height, 992/frame_width, 544/frame_height]


  1. The net outputs "detection_output" blob with shape: [1x1xNx7], where N is the number of detected pedestrians. For each detection, the description has the format: [image_id, label, conf, x_min, y_min, x_max, y_max], where:
    • image_id - ID of image in batch
    • label - ID of predicted class
    • conf - Confidence for the predicted class
    • (x_min, y_min) - Coordinates of the top left bounding box corner
    • (x_max, y_max) - Coordinates of the bottom right bounding box corner.

Legal Information

[*] Other names and brands may be claimed as the property of others.