Adaptive Distance Detection and Adaptive Instance Segmentation
Abstract
This paper is about object detection model and semantic segmentation model in an FPS game Counter Strike: Condition Zero to detect enemies. Semantic segmentation is pixel-wise detection algorithm. We customized the architecture for semantic segmentation. We used YOLOv5s to train object detection model. After training, the inference time of semantic segmentation was compared to object detection. It was observed that object detection was slightly faster and semantic segmentation is slower but better at detecting the exact shape of the object. However, we found out that this slight difference of inference can grow exponentially if the both models get more complex. To solve this, we propose a novel approach called Adaptive distance detection which combines both algorithms. We will also introduce Adaptive instance segmentation to combine semantic segmentation and panoptic segmentation.
Figures & tables
| Test accuracy | ||
|---|---|---|
| Loss function | Mean IoU | Mean F1-score |
| Jaccard | 0.70 | 0.82 |
| Dice | 0.64 | 0.78 |
| Test accuracy | ||
|---|---|---|
| Architecture | Mean IoU | Mean F1-score |
| U-Net | 0.70 | 0.82 |
| FPN | 0.63 | 0.77 |
| LinkNet | 0.45 | 0.62 |
| Comparison of inference time | |
|---|---|
| Model | Inference time (s) |
| YOLOv5s | 1.09 |
| Custom segmentation | 0.43 |
| Test accuracy | |
|---|---|
| Classifier | Mean accuracy |
| Custom | 1.00 |
| AlexNet | 1.00 |
| Test accuracy | |
|---|---|
| Classifier | Mean accuracy |
| Custom | 0.97 |
| AlexNet | 0.78 |
| Test accuracy and inference time | |||
|---|---|---|---|
| Model | Mean IoU | Mean F1-score | Mean inference (s) |
| Adaptive distance detection | 0.36 | 0.52 | 0.42 |
| Object detection (YOLOv5s) | 0.18 | 0.25 | 0.28 |
| Custom semantic segmentation | 0.56 | 0.70 | 0.48 |
| Test accuracy and inference time | |||
|---|---|---|---|
| Model | Mean IoU | Mean F1-score | Mean inference (s) |
| Adaptive instance segmentation | 0.72 | 0.83 | 1.87 |
| Panoptic segmentation | 0.75 | 0.85 | 3.83 |
| Custom semantic segmentation | 0.68 | 0.81 | 0.51 |
| Test accuracy and inference time | |||
|---|---|---|---|
| Model | Mean IoU | Mean F1-score | Mean inference (s) |
| Adaptive distance detection | 0.41 | 0.56 | 0.54 |
| Adaptive instance segmentation | 0.70 | 0.82 | 1.02 |
Appendix figures & tables7 assets
Supplementary material from the paper’s appendix.
Appendix
Object Detection
Object detection locates and classifies objects in images or video, balancing localization accuracy, robustness, and inference speed.
See all 673 papers