CReaM: Condensed Real-time Models for Depth Prediction using Convolutional Neural Networks

Since the resurgence of CNNs the robotic vision community has developed a range of algorithms that perform classification, semantic segmentation and structure prediction (depths, normals, surface curvature) using neural networks. While some of these models achieve state-of-the art results and super...

Full description

Saved in:
Bibliographic Details
Main Authors: Spek, Andrew, Dharmasiri, Thanuja, Drummond, Tom
Format: Conference Proceeding
Language:English
Subjects:
Modeling
Predictive models
Real-time systems
Robots
Semantics
Task analysis
Training
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Since the resurgence of CNNs the robotic vision community has developed a range of algorithms that perform classification, semantic segmentation and structure prediction (depths, normals, surface curvature) using neural networks. While some of these models achieve state-of-the art results and super human level performance, deploying these models in a time critical robotic environment remains an ongoing challenge. Real-time frameworks are of paramount importance to build a robotic society where humans and robots integrate seamlessly. To this end, we present a novel real-time structure prediction framework that predicts depth at 30 frames per second on an NVIDIA-TX2. At the time of writing, this is the first piece of work to showcase such a capability on a mobile platform. We also demonstrate with extensive experiments that neural networks with very large model capacities can be leveraged in order to train accurate condensed model architectures in a "from teacher to student" style knowledge transfer.
ISSN:2153-0866
DOI:10.1109/IROS.2018.8594243