A Self-supervised GAN for Unsupervised Few-shot Object Recognition

TL;DR

This paper introduces a self-supervised GAN framework for unsupervised few-shot object recognition, leveraging novel loss functions to improve classification accuracy without labeled training data.

Contribution

It extends the vanilla GAN with reconstruction and triplet losses for self-supervised learning, advancing unsupervised few-shot object recognition.

Findings

Outperforms state-of-the-art on Mini-Imagenet and Tiered-Imagenet datasets

Effective self-supervised learning with novel loss functions

Significant improvements in few-shot classification accuracy

Abstract

This paper addresses unsupervised few-shot object recognition, where all training images are unlabeled, and test images are divided into queries and a few labeled support images per object class of interest. The training and test images do not share object classes. We extend the vanilla GAN with two loss functions, both aimed at self-supervised learning. The first is a reconstruction loss that enforces the discriminator to reconstruct the probabilistically sampled latent code which has been used for generating the "fake" image. The second is a triplet loss that enforces the discriminator to output image encodings that are closer for more similar images. Evaluation, comparisons, and detailed ablation studies are done in the context of few-shot classification. Our approach significantly outperforms the state of the art on the Mini-Imagenet and Tiered-Imagenet datasets.