AttributeError: 'NoneType' object has no attribute 'set_image'

[andysingal]

Greetings, While working on creating mask using detected box, i got the following error:

  def segment(image, sam_model, boxes):
    sam_model.set_image(image)
    H, W, _ = image.shape
    boxes_xyxy = box_ops.box_cxcywh_to_xyxy(boxes) * torch.Tensor([W, H, W, H])

    transformed_boxes = sam_model.transform.apply_boxes_torch(boxes_xyxy.to(device), image.shape[:2])
    masks, _, _ = sam_model.predict_torch(
        point_coords = None,
        point_labels = None,
        boxes = transformed_boxes,
        multimask_output = False,
        )
    return masks.cpu()

  def draw_mask(mask, image, random_color=True):
      if random_color:
          color = np.concatenate([np.random.random(3), np.array([0.8])], axis=0)
      else:
          color = np.array([30/255, 144/255, 255/255, 0.6])
      h, w = mask.shape[-2:]
      mask_image = mask.reshape(h, w, 1) * color.reshape(1, 1, -1)

      annotated_frame_pil = Image.fromarray(image).convert("RGBA")
      mask_image_pil = Image.fromarray((mask_image.cpu().numpy() * 255).astype(np.uint8)).convert("RGBA")

      return np.array(Image.alpha_composite(annotated_frame_pil, mask_image_pil))

  segmented_frame_masks = segment(image_source, sam_predictor, boxes=detected_boxes)
  annotated_frame_with_mask = draw_mask(segmented_frame_masks[0][0], annotated_frame)
  Image.fromarray(annotated_frame_with_mask)

i got the error:

  ---------------------------------------------------------------------------
  AttributeError                            Traceback (most recent call last)
  [<ipython-input-74-5a61b3b98c1f>](https://localhost:8080/#) in <cell line: 1>()
  ----> 1 segmented_frame_masks = segment(image_source, sam_predictor, boxes=detected_boxes)
        2 annotated_frame_with_mask = draw_mask(segmented_frame_masks[0][0], annotated_frame)
        3 Image.fromarray(annotated_frame_with_mask)

  [<ipython-input-73-2bdf8f780a7e>](https://localhost:8080/#) in segment(image, sam_model, boxes)
       49 
       50 def segment(image, sam_model, boxes):
  ---> 51   sam_model.set_image(image)
       52   H, W, _ = image.shape
       53   boxes_xyxy = box_ops.box_cxcywh_to_xyxy(boxes) * torch.Tensor([W, H, W, H])

  AttributeError: 'NoneType' object has no attribute 'set_image'

article: https://medium.com/@amir_shakiba/sam-grounding-dino-stable-diffusion-segment-detect-change-da7926947286 colab: https://colab.research.google.com/drive/1d1zWF4YFQWwaEwYXEHEdl4jhzomZqXMb?usp=sharing

Looking forward to hearing from you Thanks

[heyoeyo]

Based on the error message, it seems like sam_model inside the segment(...) function has a value of None, and that leads to an error once you try to call the set_image(...) function on it.

In the code you provided, the sam_model is coming from the sam_predictor passed to the segment function, from this line: segmented_frame_masks = segment(image_source, sam_predictor, boxes=detected_boxes)

So that means that the sam_predictor variable is itself also None. In the colab link you provided, that variable is supposed to be set from calling the load_sam_model() function, but that function doesn't actually return anything. So to fix the problem that load_sam_model() function should probably include an extra line: return sam_model at the end of the function.

[andysingal]

Based on the error message, it seems like sam_model inside the segment(...) function has a value of None, and that leads to an error once you try to call the set_image(...) function on it.

In the code you provided, the sam_model is coming from the sam_predictor passed to the segment function, from this line: segmented_frame_masks = segment(image_source, sam_predictor, boxes=detected_boxes)

So that means that the sam_predictor variable is itself also None. In the colab link you provided, that variable is supposed to be set from calling the load_sam_model() function, but that function doesn't actually return anything. So to fix the problem that load_sam_model() function should probably include an extra line: return sam_model at the end of the function.

Thanks , let me check and get back to you. Please keep the ticket open