Fix the everything mode bug and add point mode

app.py

@@ -1,27 +1,49 @@
 from ultralytics import YOLO
 import gradio as gr
 import torch
-from tools import fast_process
+from tools import fast_process, format_results, box_prompt, point_prompt
+from PIL import ImageDraw
+import numpy as np
 
 # Load the pre-trained model
 model = YOLO('checkpoints/FastSAM.pt')
 
+device = 'cuda' if torch.cuda.is_available() else 'cpu'
+
 # Description
 title = "<center><strong><font size='8'>🏃 Fast Segment Anything 🤗</font></strong></center>"
 
-news = """ # News
+news = """ # 📖 News
 
-        🔥 Add the 'Advanced options" in Everything mode to get a more detailed adjustment.
+        🔥 2023/06/24: Add the 'Advanced options" in Everything mode to get a more detailed adjustment.
+        
+        🔥 2023/06/26: Support the points mode. (Better and faster interaction will come soon!)
+        
         """
 
          
-        # 🔥 Support the points mode and box mode, text mode will come soon.
 
-description = """This is a demo on Github project 🏃 [Fast Segment Anything Model](https://github.com/CASIA-IVA-Lab/FastSAM).
+description_e = """This is a demo on Github project 🏃 [Fast Segment Anything Model](https://github.com/CASIA-IVA-Lab/FastSAM).
                 
                 🎯 Upload an Image, segment it with Fast Segment Anything (Everything mode). The other modes will come soon.
                 
-                ⌛️ It takes about 4~ seconds to generate segment results. The concurrency_count of queue is 1, please wait for a moment when it is crowded.
+                ⌛️ It takes about 6~ seconds to generate segment results. The concurrency_count of queue is 1, please wait for a moment when it is crowded.
+                
+                🚀 To get faster results, you can use a smaller input size and leave high_visual_quality unchecked.
+                
+                📣 You can also obtain the segmentation results of any Image through this Colab: [![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/drive/1oX14f6IneGGw612WgVlAiy91UHwFAvr9?usp=sharing)
+                
+                😚 A huge thanks goes out to the @HuggingFace Team for supporting us with GPU grant.
+                
+                🏠 Check out our [Model Card 🏃](https://huggingface.co/An-619/FastSAM)
+                
+              """
+
+description_p = """This is a demo on Github project 🏃 [Fast Segment Anything Model](https://github.com/CASIA-IVA-Lab/FastSAM).
+                
+                🎯 Upload an Image, add points and segment it with Fast Segment Anything (Points mode).
+                
+                ⌛️ It takes about 6~ seconds to generate segment results. The concurrency_count of queue is 1, please wait for a moment when it is crowded.
                 
                 🚀 To get faster results, you can use a smaller input size and leave high_visual_quality unchecked.
                 
@@ -41,17 +63,14 @@ default_example = examples[0]
 css = "h1 { text-align: center } .about { text-align: justify; padding-left: 10%; padding-right: 10%; }"
 
 
-def segment_image(
+def segment_everything(
     input,
     input_size=1024, 
     iou_threshold=0.7,
     conf_threshold=0.25,
     better_quality=False,
-    mask_random_color=True,
     withContours=True,
-    points=None,
-    bbox=None,
-    point_label=None,
+    mask_random_color=True,
     use_retina=True,
     ):
     input_size = int(input_size)  # 确保 imgsz 是整数
@@ -69,19 +88,80 @@ def segment_image(
                     iou=iou_threshold,
                     conf=conf_threshold,
                     imgsz=input_size,)
+    
     fig = fast_process(annotations=results[0].masks.data,
                         image=input,
                         device=device,
                         scale=(1024 // input_size),
                         better_quality=better_quality,
                         mask_random_color=mask_random_color,
-                        points=points,
-                        bbox=bbox,
-                        point_label=point_label,
+                        bbox=None,
                         use_retina=use_retina,
                         withContours=withContours,)
     return fig
 
+def segment_with_points(
+    input,
+    input_size=1024, 
+    iou_threshold=0.7,
+    conf_threshold=0.25,
+    better_quality=False,
+    withContours=True,
+    mask_random_color=True,
+    use_retina=True,
+    ):    
+    global global_points
+    global global_point_label
+    
+    input_size = int(input_size)  # 确保 imgsz 是整数
+    # Thanks for the suggestion by hysts in HuggingFace.
+    w, h = input.size
+    scale = input_size / max(w, h)
+    new_w = int(w * scale)
+    new_h = int(h * scale)
+    input = input.resize((new_w, new_h))
+    
+    scaled_points = [[int(x * scale) for x in point] for point in global_points]
+
+    results = model(input,
+                    device=device,
+                    retina_masks=True,
+                    iou=iou_threshold,
+                    conf=conf_threshold,
+                    imgsz=input_size,)
+    
+    results = format_results(results[0], 0)
+    
+    annotations, _ = point_prompt(results, scaled_points, global_point_label, new_h, new_w)
+    annotations = np.array([annotations])
+        
+    fig = fast_process(annotations=annotations,
+                        image=input,
+                        device=device,
+                        scale=(1024 // input_size),
+                        better_quality=better_quality,
+                        mask_random_color=mask_random_color,
+                        bbox=None,
+                        use_retina=use_retina,
+                        withContours=withContours,)
+    global_points = []
+    global_point_label = []
+    return fig, None
+
+def get_points_with_draw(image, label, evt: gr.SelectData):
+    x, y = evt.index[0], evt.index[1]
+    point_radius, point_color = 15, (255, 255, 0) if label == 'Add Mask' else (255, 0, 255)
+    global global_points
+    global global_point_label
+    print((x, y))
+    global_points.append([x, y])
+    global_point_label.append(1 if label == 'Add Mask' else 0)
+    
+    # 创建一个可以在图像上绘图的对象
+    draw = ImageDraw.Draw(image)
+    draw.ellipse([(x - point_radius, y - point_radius), (x + point_radius, y + point_radius)], fill=point_color)
+    return image
+    
 
 # input_size=1024
 # high_quality_visual=True
@@ -93,75 +173,119 @@ def segment_image(
 # pil_image = fast_process(annotations=results[0].masks.data,
 #                             image=input, high_quality=high_quality_visual, device=device)
 
-device = 'cuda' if torch.cuda.is_available() else 'cpu'
+cond_img_e = gr.Image(label="Input", value=default_example[0], type='pil')
+cond_img_p = gr.Image(label="Input with points", value=default_example[0], type='pil')
 
-cond_img = gr.Image(label="Input", value=default_example[0], type='pil')
+segm_img_e = gr.Image(label="Segmented Image", interactive=False, type='pil')
+segm_img_p = gr.Image(label="Segmented Image with points", interactive=False, type='pil')
 
-segm_img = gr.Image(label="Segmented Image", interactive=False, type='pil')
+global_points = []
+global_point_label = [] # TODO:Clear points each image
 
 input_size_slider = gr.components.Slider(minimum=512,
                                          maximum=1024,
                                          value=1024,
                                          step=64,
-                                         label='Input_size (Our model was trained on a size of 1024)')
+                                         label='Input_size',
+                                         info='Our model was trained on a size of 1024')
 
 with gr.Blocks(css=css, title='Fast Segment Anything') as demo:
     with gr.Row():
-        with gr.Column(scale=1):
-            # Title
-            gr.Markdown(title)
-    
-        with gr.Column(scale=1):
-            # News
-            gr.Markdown(news)
-
-    # Images
-    with gr.Row(variant="panel"):
-        with gr.Column(scale=1):
-            cond_img.render()
-
-        with gr.Column(scale=1):
-            segm_img.render()
-
-    # Submit & Clear
-    with gr.Row():
-        with gr.Column():
-            input_size_slider.render()
-
-            with gr.Row():
-                contour_check = gr.Checkbox(value=True, label='withContours')
-
-                with gr.Column():
-                    segment_btn = gr.Button("Segment Anything", variant='primary')
-
-                # with gr.Column():
-                # clear_btn = gr.Button("Clear", variant="primary")
-
-            gr.Markdown("Try some of the examples below ⬇️")
-            gr.Examples(examples=examples,
-                        inputs=[cond_img],
-                        outputs=segm_img,
-                        fn=segment_image,
-                        cache_examples=True,
-                        examples_per_page=4)
-
-        with gr.Column():
-            with gr.Accordion("Advanced options", open=False):
-                iou_threshold = gr.Slider(0.1, 0.9, 0.7, step=0.1, label='iou_threshold')
-                conf_threshold = gr.Slider(0.1, 0.9, 0.25, step=0.05, label='conf_threshold')
-                mor_check = gr.Checkbox(value=False, label='better_visual_quality')
+            with gr.Column(scale=1):
+                # Title
+                gr.Markdown(title)
+        
+            with gr.Column(scale=1):
+                # News
+                gr.Markdown(news)
                 
-            # Description
-            gr.Markdown(description)
-
-    segment_btn.click(segment_image,
-                    inputs=[cond_img, input_size_slider, iou_threshold, conf_threshold, mor_check, contour_check],
-                    outputs=segm_img)
-
-    # def clear():
-    # return None, None
-
-    # clear_btn.click(fn=clear, inputs=None, outputs=None)
+    with gr.Tab("Everything mode"):
+        # Images
+        with gr.Row(variant="panel"):
+            with gr.Column(scale=1):
+                cond_img_e.render()
+
+            with gr.Column(scale=1):
+                segm_img_e.render()
+
+        # Submit & Clear
+        with gr.Row():
+            with gr.Column():
+                input_size_slider.render()
+
+                with gr.Row():
+                    contour_check = gr.Checkbox(value=True, label='withContours', info='draw the edges of the masks')
+
+                    with gr.Column():
+                        segment_btn_e = gr.Button("Segment Everything", variant='primary')
+                        clear_btn_e = gr.Button("Clear", variant="secondary")
+
+                gr.Markdown("Try some of the examples below ⬇️")
+                gr.Examples(examples=examples,
+                            inputs=[cond_img_e],
+                            outputs=segm_img_e,
+                            fn=segment_everything,
+                            cache_examples=True,
+                            examples_per_page=4)
+
+            with gr.Column():
+                with gr.Accordion("Advanced options", open=False):
+                    iou_threshold = gr.Slider(0.1, 0.9, 0.7, step=0.1, label='iou', info='iou threshold for filtering the annotations')
+                    conf_threshold = gr.Slider(0.1, 0.9, 0.25, step=0.05, label='conf', info='object confidence threshold')
+                    with gr.Row():
+                        mor_check = gr.Checkbox(value=False, label='better_visual_quality', info='better quality using morphologyEx')
+                        with gr.Column():
+                            retina_check = gr.Checkbox(value=True, label='use_retina', info='draw high-resolution segmentation masks')
+                    
+                # Description
+                gr.Markdown(description_e)
+
+    with gr.Tab("Points mode"):
+        # Images
+        with gr.Row(variant="panel"):
+            with gr.Column(scale=1):
+                cond_img_p.render()
+
+            with gr.Column(scale=1):
+                segm_img_p.render()
+                
+        # Submit & Clear
+        with gr.Row():
+            with gr.Column():
+                with gr.Row():
+                    add_or_remove = gr.Radio(["Add Mask", "Remove Area"], value="Add Mask", label="Point_label (foreground/background)")
+
+                    with gr.Column():
+                        segment_btn_p = gr.Button("Segment with points prompt", variant='primary')
+                        clear_btn_p = gr.Button("Clear points", variant='secondary')
+
+                gr.Markdown("Try some of the examples below ⬇️")
+                gr.Examples(examples=examples,
+                            inputs=[cond_img_p],
+                            outputs=segm_img_p,
+                            fn=segment_with_points,
+                            # cache_examples=True,
+                            examples_per_page=4)
+
+            with gr.Column():
+                # Description
+                gr.Markdown(description_p)
+        
+    cond_img_p.select(get_points_with_draw, [cond_img_p, add_or_remove], cond_img_p)
+
+    segment_btn_e.click(segment_everything,
+                    inputs=[cond_img_e, input_size_slider, iou_threshold, conf_threshold, mor_check, contour_check, retina_check],
+                    outputs=segm_img_e)
+    
+    segment_btn_p.click(segment_with_points,
+                    inputs=[cond_img_p],
+                    outputs=[segm_img_p, cond_img_p])
+    
+    def clear():
+        return None, None
+    
+    clear_btn_e.click(clear, outputs=[cond_img_e, segm_img_e])
+    clear_btn_p.click(clear, outputs=[cond_img_p, segm_img_p])
 
 demo.queue()
 demo.launch()

tools.py

@@ -93,9 +93,7 @@ def fast_process(
     scale,
     better_quality=False,
     mask_random_color=True,
-    points=None,
     bbox=None,
-    point_label=None,
     use_retina=True,
     withContours=True,
     ):
@@ -117,8 +115,6 @@ def fast_process(
             plt.gca(),
             random_color=mask_random_color,
             bbox=bbox,
-            points=points,
-            pointlabel=point_label,
             retinamask=use_retina,
             target_height=original_h,
             target_width=original_w,
@@ -131,8 +127,6 @@ def fast_process(
             plt.gca(),
             random_color=mask_random_color,
             bbox=bbox,
-            points=points,
-            pointlabel=point_label,
             retinamask=use_retina,
             target_height=original_h,
             target_width=original_w,
@@ -159,7 +153,7 @@ def fast_process(
         cv2.drawContours(temp, contour_all, -1, (255, 255, 255), 2 // scale)
         color = np.array([0 / 255, 0 / 255, 255 / 255, 0.9])
         contour_mask = temp / 255 * color.reshape(1, 1, -1)
-    i
+
     image = image.convert('RGBA')
     overlay_inner = Image.fromarray((inner_mask * 255).astype(np.uint8), 'RGBA')
     image.paste(overlay_inner, (0, 0), overlay_inner)
@@ -177,8 +171,6 @@ def fast_show_mask(
     ax,
     random_color=False,
     bbox=None,
-    points=None,
-    pointlabel=None,
     retinamask=True,
     target_height=960,
     target_width=960,
@@ -209,16 +201,6 @@ def fast_show_mask(
     if bbox is not None:
         x1, y1, x2, y2 = bbox
         ax.add_patch(plt.Rectangle((x1, y1), x2 - x1, y2 - y1, fill=False, edgecolor='b', linewidth=1))
-    # draw point
-    if points is not None:
-        plt.scatter([point[0] for i, point in enumerate(points) if pointlabel[i] == 1],
-                    [point[1] for i, point in enumerate(points) if pointlabel[i] == 1],
-                    s=20,
-                    c='y')
-        plt.scatter([point[0] for i, point in enumerate(points) if pointlabel[i] == 0],
-                    [point[1] for i, point in enumerate(points) if pointlabel[i] == 0],
-                    s=20,
-                    c='m')
 
     if retinamask == False:
         mask = cv2.resize(mask, (target_width, target_height), interpolation=cv2.INTER_NEAREST)
@@ -231,8 +213,6 @@ def fast_show_mask_gpu(
     ax,
     random_color=False,
     bbox=None,
-    points=None,
-    pointlabel=None,
     retinamask=True,
     target_height=960,
     target_width=960,
@@ -269,20 +249,6 @@ def fast_show_mask_gpu(
                 (x1, y1), x2 - x1, y2 - y1, fill=False, edgecolor="b", linewidth=1
             )
         )
-    # draw point
-    if points is not None:
-        plt.scatter(
-            [point[0] for i, point in enumerate(points) if pointlabel[i] == 1],
-            [point[1] for i, point in enumerate(points) if pointlabel[i] == 1],
-            s=20,
-            c="y",
-        )
-        plt.scatter(
-            [point[0] for i, point in enumerate(points) if pointlabel[i] == 0],
-            [point[1] for i, point in enumerate(points) if pointlabel[i] == 0],
-            s=20,
-            c="m",
-        )
     if retinamask == False:
         mask_cpu = cv2.resize(
             mask_cpu, (target_width, target_height), interpolation=cv2.INTER_NEAREST