ValueError: Could not resolve color by class becauseDetections do not have class_id

[bb6698]

Search before asking

• I have searched the Supervision issues and found no similar bug report.

Bug

Hi!! I'm using supervision for it can view the result of segment-anything following tutorial: https://blog.roboflow.com/how-to-use-segment-anything-model-sam/

The problem is when I want to draw masks on an image using provided detections with Segment-Anything (SAM).

The error is:

annotated_image = mask_annotator.annotate(scene=image_bgr.copy(), detections=detections)
File "/home/oem/.local/lib/python3.8/site-packages/supervision/annotators/core.py", line 143, in annotate
idx = resolve_color_idx(
File "/home/oem/.local/lib/python3.8/site-packages/supervision/annotators/utils.py", line 34, in resolve_color_idx
raise ValueError(
ValueError: Could not resolve color by class becauseDetections do not have class_id

Environment

• Supervision: 0.15.0
• Python: 3.8
• Ubuntu: 20.04
• python-opencv: 4.8.1.78
• Segment-Anything

Minimal Reproducible Example

import cv2
from segment_anything import sam_model_registry, SamAutomaticMaskGenerator, SamPredictor
import supervision as sv
import time

IMAGE_PATH = "image1.jpg"

if name == 'main':
sam_checkpoint = "sam_vit_h_4b8939.pth"
model_type = "default"
device = "cuda"

sam = sam_model_registry[model_type](checkpoint=sam_checkpoint)
sam.to(device=device)

mask_generator = SamAutomaticMaskGenerator(sam)

image_bgr = cv2.imread(IMAGE_PATH)
image_rgb = cv2.cvtColor(image_bgr, cv2.COLOR_BGR2RGB)


result = mask_generator.generate(image_rgb)

mask_annotator = sv.MaskAnnotator()
detections = sv.Detections.from_sam(result)
annotated_image = mask_annotator.annotate(scene=image_bgr.copy(), detections=detections)

Additional

No response

Are you willing to submit a PR?

• Yes I'd like to help by submitting a PR!

[github-actions[bot]]

Hello there, thank you for opening an Issue ! 🙏🏻 The team was notified and they will get back to you asap.

[onuralpszr]

Hello, please do this

mask_annotator = sv.MaskAnnotator(color_map = "index") should work. Default color map is "class" but for sam it doesn't have "class_id" so you need to use "index"

[onuralpszr]

@josh-001 what is your sv version ?

[josh-001]

!pip install roboflow supervision==0.15.0
this is the version i installed.

[onuralpszr]

Install latest it will fix itself.

[josh-001]

i tried installing by !pip install -q supervision still its throwing the same error.

[onuralpszr]

using command install won't update version you need to upgrade

pip install -q supervision --upgrade supervision
or

pip install supervision==0.18.0

[SkalskiP]

Hi, @bb6698 👋🏻! Please use the tip provided by @onuralpszr. I just updated our blog post to use the latest API. I'm closing the issue, but feel free to reopen it if you will face more issues.

[bb6698]

Thanks!! I probe mask_annotator = sv.MaskAnnotator(color_map = "index") and now it works.

Cheers :)

[SambitPrabhu]

In the latest update, it seems there is no argument called 'color_map' in MaskAnnotator, and neither does 'color_lookup' exist, which is given in the documentation of MaskAnnotator on Roboflow, and seems to be similar in functionality to 'color_map'

[schoonhovenrichard]

Indeed, I had this issue. A workaround is to downgrade supervision which worked for me:
pip uninstall roboflow supervision
pip install roboflow supervision==0.15.0

[warped-ch]

Using supervision version 0.16.0 this did the trick for me:
mask_annotator = sv.MaskAnnotator(color_lookup = sv.ColorLookup.INDEX)

[SkalskiP]

mask_annotator = sv.MaskAnnotator(color_lookup = sv.ColorLookup.INDEX) is the right solution! 🙏🏻

[onuralpszr]

Since this is a old discussion and answered already multiple time, for avoid conflicts I am closing and looking this discussion.

Thank you.