Pengxiao-Wang / video_toolkits

🎥 video_toolkits

This repo provides functions to read/write a video and draw keypoints.

** Defult BGR **

Installation

git clone https://github.com/Celeste-cj/video_toolkits.git
cd video_toolkits
python setup.py install

How To Use

VideoReader

from video_toolkits import VideoReader, VideoReaderCV2

video_path = 'path to video'
reader = VideoReader(video_path)            # VideoReader - pyav, only support video file/stream
# reader = VideoReader(0)                   # webcam 0   VideoReaderCV2 support
for img in reader:
    pass

VideoWriter

from video_toolkits import VideoWriter

img_seqs = []
out_path = '{VIDEO_NAME}.mp4'                     # currently support .mp4/.avi
succeed = VideoWriter.imgseq2video(img_seqs, out_path, fps=30)

Convert flv to mp4

from video_toolkits import flv2mp4

input_vid = "{INPUT_FOLDER}/{VID_NAME}.flv"
output_folder = "{OUTPUT_FOLDER}"                 # if output_folder == '', will use {INPUT_FOLDER}
succeed = flv2mp4(input_vid, out_folder=output_folder) 

Visualization

from video_toolkits import display, draw_sklts, put_text, draw_bbox
                                                                   
img = None
kpts = []                       # default (22, 4)
text = ""
bbox = []

display(img)                       # plot image with matplotlib.pyplot
img = draw_sklts(img, kpts, color=None, sklts=None)          # draw skeletons
img = put_text(img, text)                                    # put text
img = draw_bbox(img, bbox)

Easy Import

from video_toolkits.easy_import import *       # this will import os, math, cv2, plt, numpy

License

This is free and unencumbered software released into the public domain.

Anyone is free to copy, use, compile this software, either in source code form or as a compiled binary, for any purpose, commercial or non-commercial, and by any means.