Open m-kashani opened 4 years ago
Shows the images before and after the bounding boxes based on the given coordination.
[x] Created a folder for Top 5 and Top 20 images in drop box. People can comment on images. (interesting stuff)
[x] Created a folder for Top 5 below to see bad labeling as well. -> Observation is needed.
Observation is needed.
[ ] Search and write more about contaminated examples and how it's effective.
[ ] If Bounding BOX is BIG -> DRAW the text Larger.
3L040215_322.jpg 49 B_3D_L0647_127.jpg 49 3D_L0453_49.jpg 48 3D_L0622_91.jpg 48 3D_L0622_150.jpg 48 3D_L0623_38.jpg 47 B_3D_L0647_160.jpg 47 3L010215_778.jpg 47 3D_L0453_283.jpg 47 B_3D_L0647_488.jpg 46 3D_L0215_202.jpg 45 3D_L0623_59.jpg 45 A_3D_L0646_394.jpg 45 C_3D_L0648_194.jpg 44 A_3D_L0646_263.jpg 44
https://github.com/facebookresearch/detectron2/blob/master/detectron2/utils/visualizer.py def bbox(self): p = mask_util.frPyObjects(self.polygons, self.height, self.width) p = mask_util.merge(p) bbox = mask_util.toBbox(p) bbox[2] += bbox[0] bbox[3] += bbox[1] return bbox
Contaminated Examples:
Examples that are not labeled!
3D_L0622_176.jpg
I think that it's gonna work:
https://github.com/m-kashani/pycococreator/blob/master/examples/shapes/train/shapes_train2018/1001.jpeg
https://forum.image.sc/t/creating-binary-png-masks/26348
What it does?
Shows the images before and after the bounding boxes based on the given coordination.
TO Do NEXT:
[x] Created a folder for Top 5 and Top 20 images in drop box. People can comment on images. (interesting stuff)
[x] Created a folder for Top 5 below to see bad labeling as well. ->
Observation is needed.
[ ] Search and write more about contaminated examples and how it's effective.
[ ] If Bounding BOX is BIG -> DRAW the text Larger.
3L040215_322.jpg 49 B_3D_L0647_127.jpg 49 3D_L0453_49.jpg 48 3D_L0622_91.jpg 48 3D_L0622_150.jpg 48 3D_L0623_38.jpg 47 B_3D_L0647_160.jpg 47 3L010215_778.jpg 47 3D_L0453_283.jpg 47 B_3D_L0647_488.jpg 46 3D_L0215_202.jpg 45 3D_L0623_59.jpg 45 A_3D_L0646_394.jpg 45 C_3D_L0648_194.jpg 44 A_3D_L0646_263.jpg 44
Might be related later: