feature 'real time' added

app.py

@@ -210,6 +210,43 @@ def process_video(
             sink.write_frame(frame)
     return result_file_path
 
+def process_image_real_time(
+    input_image: np.ndarray,
+    categories: str = "safety helmet, safety glasses, hearing protectors, protective masks, safety shoes, protective gloves, seat belt, person, bicycle, car, motorcycle, airplane, dining table, bus, train, truck, boat, traffic light, fire hydrant, stop sign, parking meter, bench, bird, cat, dog, horse, sheep, cow, elephant, bear, zebra, giraffe, backpack, umbrella, handbag, tie, suitcase, frisbee, skis, snowboard, sports ball, kite, baseball bat, baseball glove, skateboard, surfboard, tennis racket, bottle, wine glass, cup, fork, knife, spoon, bowl, banana, apple, sandwich, orange, broccoli, carrot, hot dog, pizza, donut, cake, chair, couch, potted plant, bed, toilet, tv, laptop, mouse, remote, keyboard, cell phone, microwave, oven, toaster, sink, refrigerator, book, clock, vase, scissors, teddy bear, hair drier, toothbrush",
+    confidence_threshold: float = 0.3,
+    iou_threshold: float = 0.5,
+    # with_segmentation: bool = True,
+    with_confidence: bool = False,
+    with_class_agnostic_nms: bool = False,) -> np.ndarray:
+    # cleanup of old video files
+
+    remove_files_older_than(RESULTS, 30)
+
+    categories = process_categories(categories)
+    YOLO_WORLD_MODEL.set_classes(categories)
+    results = YOLO_WORLD_MODEL.infer(input_image, confidence=0.02)
+    detections = sv.Detections.from_inference(results)
+    detections = detections.with_nms(
+        class_agnostic=with_class_agnostic_nms,
+        threshold=iou_threshold
+    )
+    # if with_segmentation:
+    #     detections.mask = inference_with_boxes(
+    #         image=input_image,
+    #         xyxy=detections.xyxy,
+    #         model=EFFICIENT_SAM_MODEL,
+    #         device=DEVICE
+    #     )
+    output_image = cv2.cvtColor(input_image, cv2.COLOR_RGB2BGR)
+    output_image = annotate_image(
+        input_image=output_image,
+        detections=detections,
+        categories=categories,
+        with_confidence=with_confidence
+    )
+    return cv2.cvtColor(output_image, cv2.COLOR_BGR2RGB)
+
+
 
 confidence_threshold_component = gr.Slider(
     minimum=0,
@@ -265,7 +302,7 @@ with_class_agnostic_nms_component = gr.Checkbox(
 
 with gr.Blocks() as demo:
     gr.Markdown(MARKDOWN)
-    with gr.Accordion("Confiduração", open=False):
+    with gr.Accordion("Configuração", open=False):
         confidence_threshold_component.render()
         iou_threshold_component.render()
         with gr.Row():
@@ -340,6 +377,10 @@ with gr.Blocks() as demo:
         #     ],
         #     outputs=output_image_component
         # )
+    with gr.Tab(label="Tempo Real"):
+        with gr.Row():
+            input_real_time_component = gr.Image(label="Entrada", sources=["webcam"], streaming=True)
+            output_real_time_component = gr.Image(label="Saída")
 
     image_submit_button_component.click(
         fn=process_image,
@@ -367,5 +408,6 @@ with gr.Blocks() as demo:
         ],
         outputs=output_video_component
     )
+    input_real_time_component.change(fn=process_image_real_time,inputs=input_real_time_component, outputs=output_real_time_component)
 
-demo.launch(debug=False, show_error=True, max_threads=1)
+demo.launch(debug=False, show_error=True, max_threads=1, share=True)