Table 1 Comparison of representative pretrained assistive systems.
Model | Application | Inference speed | Accuracy | Resource usage | Notes |
|---|---|---|---|---|---|
YOLOv3 | Object Detection | Fast | Medium | High | Multi-object detection in real-time |
ResNet50 | Image Classification | Moderate | High | High | Accurate feature extraction |
MobileNetV2 | Embedded Assistive | Fast | Medium | Low | Suitable for Raspberry Pi and edge devices |
BLIP / ViT-GPT2 | Image Captioning | Moderate | High | Moderate | Rich semantic captions with transformer |
