# A Robotic Approach towards Quantifying Epipelagic Bound Plastic Using Deep Visual Models

arXiv: https://arxiv.org/abs/2105.01882 · HTML: https://mlsanity.com/paper/2105.01882/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 11 of its 39 references that are in the map.
- First posted on arXiv 5 May 2021.

## Builds on (most cited first, 11 of 11)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Very Deep Convolutional Networks for Large-Scale Image Recognition](https://mlsanity.com/paper/1409.1556/) (2014)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [Rich feature hierarchies for accurate object detection and semantic segmentation](https://mlsanity.com/paper/1311.2524/) (2013)
- [YOLOv4: Optimal Speed and Accuracy of Object Detection](https://mlsanity.com/paper/2004.10934/) (2020)
- [Path Aggregation Network for Instance Segmentation](https://mlsanity.com/paper/1803.01534/) (2018)
- [AutoAugment: Learning Augmentation Policies from Data](https://mlsanity.com/paper/1805.09501/) (2018)
- [CSPNet: A New Backbone that can Enhance Learning Capability of CNN](https://mlsanity.com/paper/1911.11929/) (2019)
- [DVC: An End-to-end Deep Video Compression Framework](https://mlsanity.com/paper/1812.00101/) (2018)
- [Robotic Detection of Marine Litter Using Deep Visual Detection Models](https://mlsanity.com/paper/1804.01079/) (2018)
- [Deep Neural Networks for Marine Debris Detection in Sonar Images](https://mlsanity.com/paper/1905.05241/) (2019)

## Built on by (most cited first, 1 of 1)

- [Exploiting Stability-Plasticity Asymmetry in Pretrained Detectors for Incremental Object Detection](https://mlsanity.com/paper/2504.10214/) (2025)

## Abstract (arXiv)

The quantification of positively buoyant marine plastic debris is critical to understanding how plastic litter accumulates across the world's oceans and is also crucial to identifying hotspots for targeted cleanup efforts. Currently, the most common method to quantify marine plastic is using manta trawls for manual sampling. However, this method is cost-intensive and requires human labor. This study removes the need for manual sampling by using an autonomous method using neural networks and computer vision models, which trained on images captured from various layers of the ocean column to perform real-time plastic quantification. The best performing model has a Mean Average Precision of 85% and an F1-Score of 0.89 while maintaining near real-time processing speeds ~2 ms/img.