University of Illinois Urbana-Champaign

TRAINING VISION CLASSIFIERS WITH ONE ANNOTATION PER CLASS USING SEMANTIC DATA AUGMENTATIONS

Jazra, Cyril

Content Files
SP23-ECE499-Thesis-Jazra, Cyril.pdf

Permalink

https://hdl.handle.net/2142/124844

Description

Title
TRAINING VISION CLASSIFIERS WITH ONE ANNOTATION PER CLASS USING SEMANTIC DATA AUGMENTATIONS
Author(s)
Jazra, Cyril
Issue Date
2023-05-01
Keyword(s)
Machine Learning; Computer Vision; Image Classification; Semi-supervised Learning; Self-supervised Learning; Diffusion Models.
Date of Ingest
2024-10-14T10:33:04-05:00
Abstract
In manymachinelearning applications, it is often expensive and time-consuming to collect annotated training data. A significant amount of research effort has therefore been devoted to developing techniques that enable training from as little annotated data as possible. Towards this goal, automatic data augmentations have proven to be very effective, particularly for image data. Existing forms of data augmentation often focus on geometric image transformations like cropping and color jittering. In this work, we explore a different kind of data augmentation, which we call “semantic augmentations,” and which can be added to most existing semi-supervised learning methods. New semantic augmentations of existing labeled training images are obtained by denoising a noisy version of the training data using a diffusion model. We show that by adding our semantic augmentation module to FixMatch, one of the leading semi-supervised learning methods, we obtain accuracy gains of up to 10% in the extreme case where we have only one labeled training example per class.
Type of Resource
text
Genre of Resource
dissertation/thesis
Language
eng

Owning Collections

Senior Theses - Electrical and Computer Engineering PRIMARY
The best of ECE undergraduate research