matterport / Mask_RCNN

Mask R-CNN for object detection and instance segmentation on Keras and TensorFlow
Other
24.65k stars 11.7k forks source link

no instances to display error #844

Open bellakate opened 6 years ago

bellakate commented 6 years ago

hi..I trained the model on multiple object classes.After running detection on a test image it shows no instances to display..can anyone suggest a solution please.

rbavery commented 6 years ago

I referenced this issue here: https://github.com/matterport/Mask_RCNN/issues/801 and am still waiting for a response. If this problem occurs when you are using detect_molded, you might try to see if step-by-step detection results in any predictions, this can be found at the bottom of the inspect_model notebook. I found that step by step detection shows that my model performs ok and actually predicts some masks that correspond with groundtruth, even though model.detect_molded fails.

I'm using the original MaskRCNN and still getting error with detect_molded. Any help is greatly appreciated!

preetkhaturia commented 6 years ago

Hello.. I faced the same problem. Th problem might be because of no boxes in input. Check visualise file in MRCNN folder. Thanks

1044886782 commented 5 years ago

Hello.. I faced the same problem. Th problem might be because of no boxes in input. Check visualise file in MRCNN folder. Thanks

hi, have you solved the no boxes in input?

vlzzhou commented 4 years ago

hi, have you solved the no boxes in input?

Hello.. I faced the same problem. Th problem might be because of no boxes in input. Check visualise file in MRCNN folder. Thanks

hi, have you solved the no boxes in input?