Frame metadata, describing the image dimensions, rotation, and sequencing information.
|
Frame.Metadata(Frame.Metadata
metadata)
|
| int |
getFormat()
Returns the format of this image if
image data is set.
|
| int |
getHeight()
Returns the frame height.
|
| int |
getId()
Returns the frame ID.
|
| int |
getRotation()
Returns the image rotation, indicating the counter-clockwise rotation from the
upright orientation.
|
| long |
getTimestampMillis()
Returns the timestamp, in milliseconds.
|
| int |
getWidth()
Returns the frame width.
|
Returns the format of this image if
image data is set.
NV16,
NV21,
YV12
or YUV_420_888.Returns the frame ID. A frame source such as a live video camera or a video player is expected to assign IDs in monotonically increasing order, to indicate the sequence that the frame appeared relative to other frames.
Returns the image rotation, indicating the counter-clockwise rotation from the
upright orientation. Has the same meaning as in getRotation().
Since the camera may deliver images that are rotated (e.g., if the user holds the device upside down), specifying the rotation with the image indicates how to make the image be upright, if necessary. Some detectors may rely upon rotating the image before attempting detection, whereas others may not. In preserving the original image from the camera along with this value, the detector may choose whether to make this correction (and to assume the associated cost).
However, note that the detector is expected to report detection position coordinates
that are relative to the upright version of the image (whether or not the image was
actually rotated by the detector). The Detector
will always deliver frame metadata to the Detector.Processor
that indicates the dimensions and orientation of an unrotated, upright frame.
ROTATION_0,
ROTATION_90,
ROTATION_180,
ROTATION_270.Returns the timestamp, in milliseconds.
A frame source such as a live video camera or a video player is expected to assign timestamps in a way that makes sense for the medium. For example, live video may use the capture time of each frame, whereas a video player may use the elapsed time to the frame within the video. Timestamps should be in monotonically increasing order, to indicate the passage of time.