Advancements in Data Augmentation Techniques

Organisers

Length

2 hours

Description

Rapid advances in computer vision have been driven by neural network techniques, particularly deep learning (DL) models. These models excel in image analysis and understanding, thanks to the availability of large-scale image datasets for training. However, the performance of DL models can still be limited due to a lack of diverse and representative training samples. Data augmentation techniques address this challenge by artificially expanding the training dataset through various transformations and modifications applied to existing images. By increasing variability and diversity, data augmentation enhances the model's ability to generalize and learn robust features.

This special session at the IMVIP conference focuses on data augmentation advancements, encompassing innovative methods, domain-specific approaches, automated strategies, evaluation metrics, and their influence on computer vision tasks, with a particular emphasis on promoting debiasing, fairness, and transparency in model outcomes. This special session will bring together experts to discuss the latest research findings, challenges, and opportunities in the data augmentation area.

Category/Keywords:

Potential topics include but are not limited to the following:

• Novel data augmentation techniques

• Data augmentation for audio and/or video data

• Data augmentation for text data

• Domain-specific data augmentation approaches for specialized tasks

• Automated and learned data augmentation strategies

• Evaluation metrics for assessing the effectiveness of augmented datasets

• Impact of data augmentation on object recognition and Classification

• Data augmentation for improving image segmentation algorithms

• Generating realistic synthetic data through augmentation

• Augmentation techniques for handling imbalanced datasets

• Adversarial attacks and defenses in the context of data augmentation

• Transfer learning with augmented datasets for improved performance

• Data augmentation for improving robustness to noise and occlusions

• Augmentation methods for enhancing fine-grained image analysis

• Combining multiple augmentation techniques for improved generalization

• Real-time data augmentation for efficient computer vision applications.

• Debiasing methods in data augmentation

• Fairness considerations in data augmentation

• Transparency approaches in data augmentation

• Domain-specific data augmentation strategies

• Evaluation metrics for assessing data augmentation effectiveness

• Novel dataset 

Submission Instructions

Submission to the workshop follows the same procedures as for main conference papers. This workshop supports both ordinary and short paper submissions.  please select the specific Special Session name you are interested in the "additional questions" part of the submission. Please note that submission dates are the same as per the main conference schedule.