Table 4.
CNN Case Study.
| Case Study | CNN Description | Number of Classes | Detection Accuracy | Average Detection Time |
|---|---|---|---|---|
| Garbage detection on grass [8] | SegNet + ResNet | 5 | 96 | 8.1 |
| Marine debris detection [41] | Faster RCNN Inception v2 | 3 | 81.0 | NA |
| SSD MobileNet v2 | 3 | 69.8 | ||
| Tiny-YOLO | 3 | 31.6 | ||
| Floor debris detection [13] | Faster RCNN ResNet | 2 | 97.8 | 184.1 |
| Mobilenet V2 SSD | 95.5 | 71 | ||
| Trash classification [27] | 11 layer CNN | 6 | 22 | NA |
| Proposed system | Customized 16 layer CNN | 2 | 96 | NA |