Skip to content

bug when generating mask #239

@shi-yan

Description

@shi-yan

seg = coco_kpt(i).annorect(p).segmentation{1};

segmentation is an array, but the code only processes the first element of the array

for example:

"segmentation": [
    [329.88, 211.23, 337.63, 211.93, 341.51, 210.17, 341.86, 206.65, 340.1, 202.77, 335.52, 199.6, 335.52, 197.13, 335.52, 190.44, 339.04, 186.21, 342.57, 170.35, 345.74, 159.43, 342.57, 154.14, 340.1, 148.86, 336.93, 143.92, 337.28, 138.28, 344.68, 138.28, 340.45, 122.43, 339.75, 116.79, 334.46, 107.63, 320.37, 104.1, 325.65, 97.06, 326, 91.06, 321.77, 82.25, 316.49, 89.66, 305.21, 87.19, 304.51, 92.47, 304.16, 97.06, 305.21, 100.58, 304.16, 103.4, 292.53, 105.51, 284.42, 123.84, 278.43, 140.05, 273.14, 146.74, 273.85, 152.73, 279.49, 156.26, 281.6, 154.14, 284.07, 144.63, 289, 133, 293.23, 133.7, 294.99, 125.95, 295.35, 157.67, 312.26, 157.31, 318.96, 150.27, 335.17, 150.27, 333.76, 157.31, 324.95, 158.02, 315.43, 164.01, 309.44, 170, 306.27, 176.34, 305.56, 185.51, 310.85, 184.8, 313.67, 178.81, 312.97, 170.7, 316.14, 165.42, 320.01, 162.25, 323.18, 167.89, 327.06, 172.11, 333.4, 171.06, 330.23, 178.46, 330.58, 190.09, 329.18, 197.13, 329.18, 208.76],
    [307.1, 171.37, 309.57, 164.78, 295.02, 166.7, 296.11, 173.02, 299.68, 170.54, 306, 172.19]
],

as a result, some parts are missing from the mask:

Screenshot from 2020-01-14 15-18-50

for example, left is when I process all segments, right is the result of the current code, some legs are missing.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions