how to see several images about hand-detection-tutorial HOT 6 CLOSED

I'm not sure. Could you try to modify this line of code in model_lib.py to the following? (You could replace 64 with a larger number if you'd like.)

        img_summary = tf.summary.image('Detections_Left_Groundtruth_Right',
                                       detection_and_groundtruth,
                                       max_outputs=64)

Reference: https://www.tensorflow.org/api_docs/python/tf/compat/v1/summary/image

Let me know if it works.

from hand-detection-tutorial.

It works the same as before... Thanks.

from hand-detection-tutorial.

I see. I guess we need to modify the code so that it saves all evaluated images (into a big batch) and calls tf.summary.imag() with those altogether. It would probably require a lot of code changes.

From what I can see, the original tensorflow object detection API code simply does not support that. I guess you could file this as a feature request to tensorflow/models.

from hand-detection-tutorial.

Thank you for your reply!!

from hand-detection-tutorial.

https://drive.google.com/file/d/18kzG7vUzHhlSB9hRI_0Z6Voc2j9z3WLM/view?usp=sharing

Hi, I could see these results like you when I ran the 'eval.sh'.
But I don't know what this means.
I know about "IoU", I don't know "0.50:0.95", "area", "maxDets"'s mean.

from hand-detection-tutorial.

https://jkjung-avt.github.io/trt-detection-map/

from hand-detection-tutorial.