video
video
#
video_dataset — every stride-th frame of a posed video as a posed image.
video_dataset(video, stride=10, max_dimension=1600)
#
Every stride-th frame of video as a posed image, decoded in one pass.
A kept frame carries the depth the sensor measured at it, Missing
where it measured none. A frame without intrinsics is left out.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
video
|
PosedVideo
|
The posed video. |
required |
stride
|
int
|
Keep one frame in this many, starting from the first. |
10
|
max_dimension
|
int
|
The longest image side, in pixels, as written. |
1600
|
Raises:
| Type | Description |
|---|---|
ValueError
|
If |