VISER: Visual Self-Regularization

02/07/2018
by   Hamid Izadinia, et al.
0

In this work, we propose the use of large set of unlabeled images as a source of regularization data for learning robust visual representation. Given a visual model trained by a labeled dataset in a supervised fashion, we augment our training samples by incorporating large number of unlabeled data and train a semi-supervised model. We demonstrate that our proposed learning approach leverages an abundance of unlabeled images and boosts the visual recognition performance which alleviates the need to rely on large labeled datasets for learning robust representation. To increment the number of image instances needed to learn robust visual models in our approach, each labeled image propagates its label to its nearest unlabeled image instances. These retrieved unlabeled images serve as local perturbations of each labeled image to perform Visual Self-Regularization (VISER). To retrieve such visual self regularizers, we compute the cosine similarity in a semantic space defined by the penultimate layer in a fully convolutional neural network. We use the publicly available Yahoo Flickr Creative Commons 100M dataset as the source of our unlabeled image set and propose a distributed approximate nearest neighbor algorithm to make retrieval practical at that scale. Using the labeled instances and their regularizer samples we show that we significantly improve object categorization and localization performance on the MS COCO and Visual Genome datasets where objects appear in context.

READ FULL TEXT

page 2

page 3

page 5

page 6

page 9

research
09/02/2021

Semi-Supervised Learning using Siamese Networks

Neural networks have been successfully used as classification models yie...
research
06/10/2015

LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loop

While there has been remarkable progress in the performance of visual re...
research
09/04/2018

Modeling Surface Appearance from a Single Photograph using Self-augmented Convolutional Neural Networks

We present a convolutional neural network (CNN) based solution for model...
research
06/07/2022

Self-Training of Handwritten Word Recognition for Synthetic-to-Real Adaptation

Performances of Handwritten Text Recognition (HTR) models are largely de...
research
12/11/2019

Identifying Mislabeled Instances in Classification Datasets

A key requirement for supervised machine learning is labeled training da...
research
12/20/2014

Self-informed neural network structure learning

We study the problem of large scale, multi-label visual recognition with...
research
07/15/2020

How to trust unlabeled data? Instance Credibility Inference for Few-Shot Learning

Deep learning based models have excelled in many computer vision task an...

Please sign up or login with your details

Forgot password? Click here to reset