new YoloV8 Models.

app.py

@@ -1,32 +1,34 @@
 import gradio as gr
+from ultralytics import YOLO
 import numpy as np
 from PIL import Image, ImageDraw, ImageFilter, ImageOps
 from transformers import OneFormerProcessor, OneFormerForUniversalSegmentation, pipeline
 import torchvision.transforms
 import torch
 
-
-# pipeline = pipeline(task="image-segmentation", model="shi-labs/oneformer_cityscapes_swin_large", label_ids_to_fuse=[11])
-
-person_processor = OneFormerProcessor.from_pretrained("shi-labs/oneformer_cityscapes_swin_large")
-person_model = OneFormerForUniversalSegmentation.from_pretrained("shi-labs/oneformer_cityscapes_swin_large")
 transform = torchvision.transforms.ToPILImage()
+seg_model = YOLO("yolov8m-seg.pt")
+lp_model = YOLO("yolov8m_lp.pt")
 
 def detect_person(image: Image):
-    semantic_inputs = person_processor(images=image, task_inputs=["semantic"], return_tensors="pt")
-    semantic_outputs = person_model(**semantic_inputs)
-    predicted_semantic_map = person_processor.post_process_semantic_segmentation(semantic_outputs, target_sizes=[image.size[::-1]])[0]
-    mask = transform(predicted_semantic_map.to(torch.uint8))
-    mask = Image.eval(mask, lambda x: 0 if x == 11 else 255)
-    return mask  
-
-
-# def detect_person_pipeline():
-#     results = pipeline(image, label_ids_to_fuse=[11])
-#     for detection in results:
-#         if detection["label"] == "person":
-#             return ImageOps.invert(detection["mask"])
-#     return Image.new(mode="L", size=image.size, color=255)
+    result = seg_model(image, device="CPU")[0]
+    masks = result.masks.data
+    clss = result.boxes.cls
+    person_indices = torch.where(clss == 0)
+    person_masks = masks[person_indices]
+    people_mask = torch.any(person_masks, dim=0).to(torch.uint8) * 255
+    mask = transform(~people_mask)
+    mask = mask.resize((image.width, image.height), resample=Image.Resampling.BILINEAR)
+    return mask
+
+def detect_license_plate(image: Image):
+    result = lp_model(image, imgsz=(image.height, image.width), device="cpu")[0]
+    boxes = result.boxes.data[:, :4]
+    mask = Image.new(mode="L", size=image.size, color=255)
+    draw = ImageDraw.Draw(mask)
+    for box in boxes:
+        draw.rectangle(list(box), fill=0)
+    return mask
 
 
 def detect_dummy(image: Image):
@@ -35,40 +37,37 @@ def detect_dummy(image: Image):
 
 detectors = {
     "Person": detect_person,
-    # "License Plate": detect_license_plate
+    "License Plate": detect_license_plate
 }
 
 def test_gradio(image):
-    masks = [detect_person(image)]
+    masks = [detect_person(image), detect_license_plate(image)]
     combined = np.minimum.reduce([np.array(m) for m in masks])
     mask = Image.fromarray(combined)
     # Apply blur through mask
-    blurred = image.filter(ImageFilter.GaussianBlur(15))
+    blurred = image.filter(ImageFilter.GaussianBlur(30))
     anonymized = Image.composite(image, blurred, mask)
-    return anonymized
+    ## TODO: Tempfile statt einem generischen File
+    anonymized.save("anon.JPG")
+    return "anon.JPG"
 
-
-demo_live = gr.Interface(
-    fn=test_gradio, 
-    inputs=gr.Image(source="webcam", type="pil", shape=(640, 480)), 
-    outputs=gr.Image(type="pil")
-)
+# demo_live = gr.Interface(
+#     fn=test_gradio, 
+#     inputs=gr.Image(source="webcam", type="pil", shape=(640, 480)), 
+#     outputs=gr.Image(type="pil")
+# )
 
 demo_upload = gr.Interface(
     fn=test_gradio, 
-    inputs=gr.Image(type="pil", shape=(640, 480)), 
-    outputs=gr.Image(type="pil")
-)
-
-demo = gr.TabbedInterface(
-    interface_list=[demo_live, demo_upload], 
-    tab_names=["Webcam", "Bild hochladen"], 
-    title="Image Anonymizer"
+    inputs=gr.Image(type="pil"), 
+    outputs=gr.Image()
 )
 
-# if __name__ == "__main__":
-    # demo.launch()
-    # 
+# demo = gr.TabbedInterface(
+#     interface_list=[demo_live, demo_upload], 
+#     tab_names=["Webcam", "Bild hochladen"], 
+#     title="Image Anonymizer"
+# )
 
 print(__name__)
-demo.launch(server_name="localhost", server_port=8080)
+demo_upload.launch(server_name="localhost", server_port=8080)

requirements.txt

@@ -1,5 +1,5 @@
-yolov5
-transformers
+gradio
+ultralytics
 Pillow
 numpy
 torch

yolov8m-seg.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9fc740ca0824e14b44681d491dc601efa664ec6ecea9a870acf876053826448
+size 54899779

yolov8m_lp.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e57cd2006a46470b779990048e82cf1b29b2130d989a1ecdc5c610d48f19c8ca
+size 52033983