Summary of Eventaug: Multifaceted Spatio-temporal Data Augmentation Methods For Event-based Learning, by Yukun Tian et al.
EventAug: Multifaceted Spatio-Temporal Data Augmentation Methods for Event-based Learning
by Yukun Tian, Hao Chen, Yongjian Deng, Feihong Shen, Kepan Liu, Wei You, Ziyang Zhang
First submitted to arxiv on: 18 Sep 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper introduces EventAug, a systematic data augmentation scheme to enrich spatial-temporal diversity in event cameras. The community faces challenges such as over-fitting and inadequate feature learning due to limited diversity and data deficiency. To address this gap, the authors propose Multi-scale Temporal Integration (MSTI) to diversify motion speed, Spatial-salient Event Mask (SSEM), and Temporal-salient Event Mask (TSEM) to enrich object variants. The augmentation method facilitates models learning richer motion patterns, object variants, and local spatio-temporal relations, improving model robustness. Experiment results show significant improvements in accuracy on the DVS128 Gesture task. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary Event cameras are special devices that can record events as they happen, which is useful for many applications. However, the data used to train these cameras is often limited and lacks diversity, making it hard for computers to learn from them. To solve this problem, scientists have developed a new way to add variety to the data by changing the motion of objects in the videos and masking certain parts of the video. This helps computers learn better features from the data and make more accurate predictions. The results show that this method can improve accuracy by 4.87% on some tasks. |
Keywords
» Artificial intelligence » Data augmentation » Mask
