Spaces:
Running
Running
Upload 2 files
Browse files- FootDetection.py +71 -0
- requirements.txt +4 -0
FootDetection.py
ADDED
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
import torch
|
3 |
+
import torchvision
|
4 |
+
from torchvision.models.detection.faster_rcnn import FastRCNNPredictor
|
5 |
+
from PIL import Image, ImageDraw
|
6 |
+
from torchvision.transforms import functional as F
|
7 |
+
from huggingface_hub import hf_hub_download
|
8 |
+
|
9 |
+
|
10 |
+
class FootDetection:
|
11 |
+
def __init__(self, device="cpu"):
|
12 |
+
self.device = torch.device(device)
|
13 |
+
self.checkpoint_dir = "checkpoints"
|
14 |
+
self.checkpoint_file = "fasterrcnn_foot.pth"
|
15 |
+
self.model = self._load_model()
|
16 |
+
self.last_detection = None
|
17 |
+
|
18 |
+
def _load_model(self):
|
19 |
+
local_path = os.path.join(self.checkpoint_dir, self.checkpoint_file)
|
20 |
+
|
21 |
+
# Download if not exists
|
22 |
+
if not os.path.exists(local_path):
|
23 |
+
os.makedirs(self.checkpoint_dir, exist_ok=True)
|
24 |
+
print("Downloading model from Hugging Face...")
|
25 |
+
local_path = hf_hub_download(
|
26 |
+
repo_id="tonyassi/foot-detection",
|
27 |
+
filename=self.checkpoint_file,
|
28 |
+
local_dir=self.checkpoint_dir
|
29 |
+
)
|
30 |
+
|
31 |
+
# Load model
|
32 |
+
model = torchvision.models.detection.fasterrcnn_resnet50_fpn(weights="DEFAULT")
|
33 |
+
in_features = model.roi_heads.box_predictor.cls_score.in_features
|
34 |
+
model.roi_heads.box_predictor = FastRCNNPredictor(in_features, 2)
|
35 |
+
model.load_state_dict(torch.load(local_path, map_location=self.device))
|
36 |
+
model.to(self.device)
|
37 |
+
model.eval()
|
38 |
+
return model
|
39 |
+
|
40 |
+
def detect(self, image, threshold=0.1):
|
41 |
+
"""Run foot detection on a PIL image"""
|
42 |
+
image_tensor = F.to_tensor(image).unsqueeze(0).to(self.device)
|
43 |
+
with torch.no_grad():
|
44 |
+
outputs = self.model(image_tensor)[0]
|
45 |
+
|
46 |
+
boxes = []
|
47 |
+
scores = []
|
48 |
+
for box, score in zip(outputs["boxes"], outputs["scores"]):
|
49 |
+
if score >= threshold:
|
50 |
+
boxes.append(box.tolist())
|
51 |
+
scores.append(score.item())
|
52 |
+
|
53 |
+
self.last_detection = {
|
54 |
+
"boxes": boxes,
|
55 |
+
"scores": scores
|
56 |
+
}
|
57 |
+
return self.last_detection
|
58 |
+
|
59 |
+
def draw_boxes(self, image):
|
60 |
+
"""Draw the most recent detection boxes on a copy of the image"""
|
61 |
+
if self.last_detection is None:
|
62 |
+
raise ValueError("No detection results found. Run .detect(image) first.")
|
63 |
+
image_copy = image.copy()
|
64 |
+
draw = ImageDraw.Draw(image_copy)
|
65 |
+
|
66 |
+
for box, score in zip(self.last_detection["boxes"], self.last_detection["scores"]):
|
67 |
+
x0, y0, x1, y1 = box
|
68 |
+
draw.rectangle([x0, y0, x1, y1], outline="red", width=3)
|
69 |
+
draw.text((x0, y0), f"{score:.2f}", fill="red")
|
70 |
+
|
71 |
+
return image_copy
|
requirements.txt
ADDED
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
1 |
+
torch
|
2 |
+
torchvision
|
3 |
+
pillow
|
4 |
+
huggingface_hub
|