Reading: StairNet — Top-Down Semantic Aggregation (Object Detection)

Original article was published by Sik-Ho Tsang on Artificial Intelligence on Medium


Reading: StairNet — Top-Down Semantic Aggregation (Object Detection)

In this story, “StairNet: Top-Down Semantic Aggregation for Accurate One Shot Detection” (StairNet), by KAIST, is shortly presented.

  • One-stage detectors have difficulty in detecting small objects while they are competitive with two-stage methods on large objects since the lower layer that is responsible for small objects lacks strong semantics.
  • In this paper, by introducing a feature combining module that spreads out the strong semantics in a top-down manner, the final model StairNet detector unifies the multi-scale representations and semantic distribution effectively.

This is a paper in 2018 WAVC with over 40 citations. (Sik-Ho Tsang @ Medium)