|
--- |
|
language: |
|
- "en" |
|
thumbnail: "https://example.com/path/to/your/thumbnail.jpg" |
|
tags: |
|
- yolo |
|
- object-detection |
|
- image-segmentation |
|
- computer-vision |
|
- human-body-parts |
|
license: "mit" |
|
datasets: |
|
- custom_human_body_parts_dataset |
|
metrics: |
|
- mean_average_precision |
|
- intersection_over_union |
|
base_model: "ultralytics/yolov5yolov8x-seg" |
|
--- |
|
|
|
|
|
|
|
This model is a fine-tuned version of YOLOv5 for segmenting human body parts and objects. It can detect and segment 11 different classes including various body parts, outfits, and phones. |
|
|
|
|
|
|
|
- **Model Type:** YOLOv8 for Instance Segmentation |
|
- **Task:** Segmentation |
|
- **Fine-tuning Dataset:** Custom dataset of human body parts and objects |
|
- **Number of Classes:** 11 |
|
|
|
|
|
|
|
The model can detect and segment the following classes: |
|
|
|
0. Hair |
|
1. Face |
|
2. Neck |
|
3. Arm |
|
4. Hand |
|
5. Back |
|
6. Leg |
|
7. Foot |
|
8. Outfit |
|
9. Person |
|
10. Phone |
|
|
|
|
|
|
|
This model can be used for various applications, including: |
|
|
|
- Human pose estimation |
|
- Gesture recognition |
|
- Fashion analysis |
|
- Person tracking |
|
- Human-computer interaction |
|
|
|
For detailed usage instructions, please refer to the model's README file. |
|
|
|
|
|
|
|
The model was fine-tuned on a custom dataset of annotated images containing human body parts and objects. The training process involved transfer learning from the base YOLOv8 model, with adjustments made to the final layers to accommodate the new class structure. |
|
|
|
|
|
|
|
(Note: Replace these placeholder metrics with your actual evaluation results) |
|
|
|
lr/pg0:0.000572628 |
|
lr/pg1:0.000572628 |
|
lr/pg2:0.000572628 |
|
metrics/mAP50-95(B):0.53001 |
|
metrics/mAP50-95(M):0.42367 |
|
metrics/mAP50(B):0.69407 |
|
metrics/mAP50(M):0.61714 |
|
metrics/precision(B):0.7047 |
|
metrics/precision(M):0.68041 |
|
metrics/recall(B):0.68802 |
|
metrics/recall(M):0.62248 |
|
model/GFLOPs:344.557 |
|
model/parameters:71,761,441 |
|
model/speed_PyTorch(ms):5.813 |
|
train/box_loss:0.54718 |
|
train/cls_loss:0.52977 |
|
train/dfl_loss:0.95171 |
|
train/seg_loss:1.34628 |
|
val/box_loss:0.80538 |
|
val/cls_loss:0.83434 |
|
val/dfl_loss:1.18352 |
|
val/seg_loss:2.19488 |
|
|
|
|
|
|
|
|
|
- The model's performance may vary depending on lighting conditions and image quality. |
|
- It may have difficulty with occluded or partially visible body parts. |
|
- The model's performance on diverse body types and skin tones should be carefully evaluated to ensure fairness and inclusivity. |
|
|
|
|
|
|
|
Users of this model should be aware of privacy concerns related to human body detection and ensure they have appropriate consent for its application. The model should not be used for surveillance or any application that could infringe on personal privacy without explicit consent. |