[Paper Review] A Survey on Deep Learning Methods for Robot Vision

Javier Ruiz‐del‐Solar, Patricio Loncomilla+1 more|arXiv (Cornell University)|Mar 28, 2018

Advanced Neural Network Applications140 references50 citations

TL;DR

This survey reviews how deep learning is applied to robot vision, summarizes common neural models and methods, and discusses standard design tools, principal works, and future Trends in the field.

ABSTRACT

Deep learning has allowed a paradigm shift in pattern recognition, from using hand-crafted features together with statistical classifiers to using general-purpose learning procedures for learning data-driven representations, features, and classifiers together. The application of this new paradigm has been particularly successful in computer vision, in which the development of deep learning methods for vision applications has become a hot research topic. Given that deep learning has already attracted the attention of the robot vision community, the main purpose of this survey is to address the use of deep learning in robot vision. To achieve this, a comprehensive overview of deep learning and its usage in computer vision is given, that includes a description of the most frequently used neural models and their main application areas. Then, the standard methodology and tools used for designing deep-learning based vision systems are presented. Afterwards, a review of the principal work using deep learning in robot vision is presented, as well as current and future trends related to the use of deep learning in robotics. This survey is intended to be a guide for the developers of robot vision systems.

Motivation & Objective

• Motivate the study by highlighting the shift to data-driven representations in vision and its relevance to robot vision.
• Provide a comprehensive overview of deep learning concepts and neural architectures used in vision tasks.
• Describe standard methodologies and tools for building deep-learning–based vision systems for robotics.
• Review principal works applying deep learning to robot vision and discuss current and future trends.
• Offer guidance for developers designing robot-vision systems using deep learning.

Proposed method

• Describe overview of deep learning and its adoption in computer vision.
• Summarize frequently used neural models and their main application areas.
• Present standard methodology and tools for designing DL-based vision systems.
• Review principal works applying DL to robot vision and categorize by task.
• Discuss current trends and future directions in applying DL to robotics vision.

Experimental results

Research questions

• RQ1 What are the commonly used deep learning models and architectures in robot vision?
• RQ2 What standard methodologies and tools are used to design deep-learning–based vision systems in robotics?
• RQ3 What principal works exist in applying deep learning to robot vision and how are they categorized by task?
• RQ4 What are the current trends and future directions for deep learning in robot vision?

Key findings

• Deep learning has shifted robot vision toward data-driven representations and end-to-end learning.
• The survey catalogs neural models and their main application areas in vision.
• Standard design methodologies and toolchains for DL-based vision systems in robotics are identified.
• A compilation of principal works in robot vision using deep learning is provided and organized by task.
• Emerging trends and future directions in DL for robotics are discussed to guide researchers and developers.