DeepAI AI Chat
Log In Sign Up

Learning to See by Moving

by   Pulkit Agrawal, et al.

The dominant paradigm for feature learning in computer vision relies on training neural networks for the task of object recognition using millions of hand labelled images. Is it possible to learn useful features for a diverse set of visual tasks using any other form of supervision? In biology, living organisms developed the ability of visual perception for the purpose of moving and acting in the world. Drawing inspiration from this observation, in this work we investigate if the awareness of egomotion can be used as a supervisory signal for feature learning. As opposed to the knowledge of class labels, information about egomotion is freely available to mobile agents. We show that given the same number of training images, features learnt using egomotion as supervision compare favourably to features learnt using class-label as supervision on visual tasks of scene recognition, object recognition, visual odometry and keypoint matching.


page 3

page 5

page 10


Using Motion and Internal Supervision in Object Recognition

In this thesis we address two related aspects of visual object recogniti...

Feature Learning for Accelerometer based Gait Recognition

Recent advances in pattern matching, such as speech or object recognitio...

Active Object Manipulation Facilitates Visual Object Learning: An Egocentric Vision Study

Inspired by the remarkable ability of the infant visual learning system,...

Domain Generalization for Object Recognition with Multi-task Autoencoders

The problem of domain generalization is to take knowledge acquired from ...