Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Add reference on litter detection and classification #1

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
Open

Add reference on litter detection and classification #1

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
12 changes: 12 additions & 0 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,6 +30,7 @@ Feel free to add **issue** with short description of new publication or create a
|[Proença, P.F. et al., 2020](https://arxiv.org/pdf/2003.06975.pdf)| [TACO](http://tacodataset.org/) | 60, but model was tested on 10, and 1 | segmentation | Mask RCNN |1-class mAP = 15.9% <br>10-class mAP = 17.6%| [Github](https://github.com/pedropro/TACO) |
|[Wang, T. et al., 2020](https://www.mdpi.com/1424-8220/20/14/3816)| [MJU-Waste](https://github.com/realwecan/mju-waste/) | 1 | segmentation | FCN, PSPNet, CCNet, DeepLabv3 | TACO mPP - 96.07%<br>MJU-Waste mPP = 97.14% | :x: |
| [Hong, J. et al., 2020](https://arxiv.org/pdf/2007.08097.pdf) | [TrashCan 1.0](https://conservancy.umn.edu/handle/11299/214865) | 4 | segmentation, detection | Mask RCNN, Faster RCNN | Faster RCNN mAP=34.5, Mask R-CNN mAP=30.0 | :x: |
| [Ping, P. et al., 2020](https://www.worldscientific.com/doi/abs/10.1142/S0218194020400045) | self-created | 11 | detection, classification | Faster RCNN | 11-class mAP = 24% | :x: |

## Papers

Expand Down Expand Up @@ -98,6 +99,17 @@ Feel free to add **issue** with short description of new publication or create a
- algorithm: YOLO TrashNet is used. In the experiment, authors train the last 53 layers of YOLOv3 on their dataset.
- results: The mAP@50 was used as a metric to evaluate the error made by the network on unseen examples, and it reaches 59.57% (Garbage Bag - 40.56%, Garbage Dumpster - 64.57%, Garbage Bin - 65.01% and Blob - 68.15%).
- code: :x:
- Ping, P et al., Smart Street Litter Detection and Classication Based on Faster R-CNN and Edge Computing, International Journal of Software Engineering
and Knowledge Engineering, 2020. [[`pdf`](https://www.worldscientific.com/doi/abs/10.1142/S0218194020400045)]
- dataset: This paper construct a new dataset with 7059 images containing eleven classes: litter, yard debris, yard pile, tree branches, cup, bottle, tin can, plastic bag, pothole, manhole, oil stain and 5287 negative samples without any waste. The collection of the images of the dataset has been made using 5 different sources, Google Maps Street View, ImageNet, San Jose steet images collected by students, Google Images Download and Open Image Dataset.
- algorithm: Faster RCNN is used. Two study cases are proposed in the paper, an experiment on 6 classes and another on 11 classes but the model is only extensively tested on the 11 classes case.
- results:

|mAP | mAP<sub>S</sub> | mAP<sub>M</sub> | mAP<sub>L</sub> | mAP<sub>50</sub> | mAP<sub>75</sub> | AR@1 | AR@10 | AR@100 | AR@100<sub>S</sub> | AR@100<sub>M</sub> | AR@100<sub>L</sub> |
|:---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: | :---: |
| 24.0 | ≈ 8.0 | ≈ 14.0 | ≈ 32.0 | ≈ 44.0 | ≈ 22.0 | ≈ 27.0 | ≈ 45.0 | ≈ 49.0 | 24.0 | 42.0 | ≈ 54.0 |

- code: :x:

### Segmentation

Expand Down