Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Autoencoder Based Face Verification System

Enoch Solomon, Abraham Woubie, Eyael Solomon Emiru

TL;DR

This work tackles the data-labeling bottleneck in face verification by combining unsupervised autoencoder pretraining with limited-label supervised fine-tuning. The method learns embeddings from unlabeled CelebA data via an autoencoder, then initializes a DNN classifier with those weights to produce discriminative face embeddings evaluated with cosine similarity on LFW and YTF. Results show competitive performance with significantly less labeled data, and faster convergence compared to training from scratch, demonstrating a practical data-efficient approach for face verification. The findings highlight the potential of hybrid autoencoder–DNN pipelines to enable scalable, label-efficient biometric verification in real-world settings.

Abstract

The primary objective of this work is to present an alternative approach aimed at reducing the dependency on labeled data. Our proposed method involves utilizing autoencoder pre-training within a face image recognition task with two step processes. Initially, an autoencoder is trained in an unsupervised manner using a substantial amount of unlabeled training dataset. Subsequently, a deep learning model is trained with initialized parameters from the pre-trained autoencoder. This deep learning training process is conducted in a supervised manner, employing relatively limited labeled training dataset. During evaluation phase, face image embeddings is generated as the output of deep neural network layer. Our training is executed on the CelebA dataset, while evaluation is performed using benchmark face recognition datasets such as Labeled Faces in the Wild (LFW) and YouTube Faces (YTF). Experimental results demonstrate that by initializing the deep neural network with pre-trained autoencoder parameters achieve comparable results to state-of-the-art methods.

Autoencoder Based Face Verification System

TL;DR

This work tackles the data-labeling bottleneck in face verification by combining unsupervised autoencoder pretraining with limited-label supervised fine-tuning. The method learns embeddings from unlabeled CelebA data via an autoencoder, then initializes a DNN classifier with those weights to produce discriminative face embeddings evaluated with cosine similarity on LFW and YTF. Results show competitive performance with significantly less labeled data, and faster convergence compared to training from scratch, demonstrating a practical data-efficient approach for face verification. The findings highlight the potential of hybrid autoencoder–DNN pipelines to enable scalable, label-efficient biometric verification in real-world settings.

Abstract

The primary objective of this work is to present an alternative approach aimed at reducing the dependency on labeled data. Our proposed method involves utilizing autoencoder pre-training within a face image recognition task with two step processes. Initially, an autoencoder is trained in an unsupervised manner using a substantial amount of unlabeled training dataset. Subsequently, a deep learning model is trained with initialized parameters from the pre-trained autoencoder. This deep learning training process is conducted in a supervised manner, employing relatively limited labeled training dataset. During evaluation phase, face image embeddings is generated as the output of deep neural network layer. Our training is executed on the CelebA dataset, while evaluation is performed using benchmark face recognition datasets such as Labeled Faces in the Wild (LFW) and YouTube Faces (YTF). Experimental results demonstrate that by initializing the deep neural network with pre-trained autoencoder parameters achieve comparable results to state-of-the-art methods.
Paper Structure (9 sections, 3 figures, 2 tables)

This paper contains 9 sections, 3 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Proposed Method
  3. Experimental Setup and Dataset
  4. Experimental Setup
  5. Datasets
  6. Results
  7. Results on LFW
  8. Results on YTF
  9. Conclusion

Figures (3)

  • Figure 1: The proposed Face embedding extraction from autoencoder pre-training.
  • Figure 2: The left side is an autoencoder pre-training whereas the right side is with a DNN training.
  • Figure 3: The most left one is the raw face images, the middle one is the conventional DNN face image embeddings and the right most one is the proposed face image embeddings.