product-detection-0001

Use Case and High-Level Description

A product detector based on the SSD-lite architecture with MobileNetV2 as a backbone for self-checkout points of sale-related scenes. The network can detect 12 classes of objects (sprite, kool-aid, extra, ocelo, finish, mtn_dew, best_foods, gatorade, heinz, ruffles, pringles, del_monte). Labels 0 and 1 are related to background_label and undefined correspondingly.

Example

Specification

Metric Value
Average Precision (AP) @[ IoU=0.50:0.95, area=all, maxDets=100 ] 0.715
GFlops 3.598
MParams 3.212
Source framework PyTorch*

Inputs

Image, name: input, shape: 1, 3, 512, 512 in the format B, C, H, W, where:

  • B - batch size
  • C - number of channels
  • H - image height
  • W - image width

Expected color order: BGR.

Outputs

The net outputs a blob with shape: 1, 1, 200, 7 in the format 1, 1, N, 7, where N is the number of detected bounding boxes. For each detection, the description has the [image_id, label, conf, x_min, y_min, x_max, y_max] format, where:

  • image_id - ID of the image in the batch
  • label - predicted class ID
  • 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.