Distilling Knowledge From a Deep Pose Regressor Network
This paper presents a novel method to distill knowledge from a deep pose regressor network for efficient Visual Odometry (VO). Standard distillation relies on "dark knowledge" for successful knowledge transfer. As this knowledge is not available in pose regression and the teacher predictio...
Gespeichert in:
Hauptverfasser: | , , , , |
---|---|
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | This paper presents a novel method to distill knowledge from a deep pose
regressor network for efficient Visual Odometry (VO). Standard distillation
relies on "dark knowledge" for successful knowledge transfer. As this knowledge
is not available in pose regression and the teacher prediction is not always
accurate, we propose to emphasize the knowledge transfer only when we trust the
teacher. We achieve this by using teacher loss as a confidence score which
places variable relative importance on the teacher prediction. We inject this
confidence score to the main training task via Attentive Imitation Loss (AIL)
and when learning the intermediate representation of the teacher through
Attentive Hint Training (AHT) approach. To the best of our knowledge, this is
the first work which successfully distill the knowledge from a deep pose
regression network. Our evaluation on the KITTI and Malaga dataset shows that
we can keep the student prediction close to the teacher with up to 92.95%
parameter reduction and 2.12x faster in computation time. |
---|---|
DOI: | 10.48550/arxiv.1908.00858 |