Video

This class is mainly used to extract images from videos.

The project isn’t pivoting off of video data yet, so it lacks functions.

Grabbing Frames

We can grab frames from a video using anything iterable, like np.linspace.

This grabs the 0s, 50s, 100s frames

from frmodel.base.D2.video2D import Video2D
import numpy as np
from frmodel.base.consts import CONSTS
vid = Video2D.from_video("sample.mp4")

frames = vid.to_frames(np.linspace(0, 100000, 3))

Caveats

The class doesn’t provide a way to calculate the length of the video.

Grabbing Frames

class frmodel.base.D2.video2D.Video2D(vid: cv2.VideoCapture)

Bases: object

This class holds the data as OpenCV2.VideoCapture.

Extract the actual Capture with .vid property

static from_video(file_path: str)frmodel.base.D2.video2D.Video2D

Creates an instance using the file path.

to_frames(offsets_msec: List[int], failure_default: None = None)List[frmodel.base.D2.frame2D.Frame2D]

Extracts images from the video.

Returns frmodel.Image

Parameters
  • offsets_msec – The timestamps to extract images from.

  • failure_default – Value to default to on failure on offset read.

Returns

List of Images