Deprecated! Please use mmcv instead.

Introduction

cvbase is a miscellaneous set of tools which maybe helpful for computer vision research. It comprises the following parts.

IO helpers
Image/Video operations
OpenCV wrappers for python2/3 and opencv 2/3
Timer
Progress visualization
Plotting tools
Object detection utils

Try and start with

pip install cvbase

See documentation for more features and usage.

Some popular features

There are some popular features such as progress visualization, timer, video to frames/frames to videos.

Progress visualization

If you want to apply a method to a list of items and track the progress, track_progress is a good choice. It will display a progress bar to tell the progress and ETA.

import cvbase as cvb

def func(item):
    # do something
    pass

tasks = [item_1, item_2, ..., item_n]

cvb.track_progress(func, tasks)

The output is like the following.

There is another method track_parallel_progress, which wraps multiprocessing and progress visualization.

import cvbase as cvb

def func(item):
    # do something
    pass

tasks = [item_1, item_2, ..., item_n]

cvb.track_parallel_progress(func, tasks, 8)
# 8 workers

Timer

It is convinient to computer the runtime of a code block with Timer.

import time

with cvb.Timer():
    # simulate some code block
    time.sleep(1)

Or try a more flexible way.

timer = cvb.Timer()
# code block 1 here
print(timer.since_start())
# code block 2 here
print(timer.since_last_check())
print(timer.since_start())

Video/Frames conversion

To split a video into frames.

video = cvb.VideoReader('video_file.mp4')
video.cvt2frames('frame_dir')

Besides cvt2frames, VideoReader wraps many other useful methods to operate a video like a list object, like

video = cvb.VideoReader('video_file.mp4')
len(video)  # get total frame number
video[5]  # get the 6th frame
for img in video:  # iterate over all frames
    print(img.shape)

To generate a video from frames, use the frames2video method.

video = cvb.frames2video('frame_dir', 'out_video_file.avi', fps=30)

Video editing (needs ffmpeg)

To cut a video.

cvb.cut_video('input.mp4', 'output.mp4', start=3, end=10)

To join two video clips.

cvb.concat_video(['clip1.mp4', 'clip2.mp4'], 'output.mp4')

To resize a video.

cvb.resize_video('input.mp4', 'resized.mp4', (360, 240))
# or
cvb.resize_video('input.mp4', 'resized.mp4', ratio=2)

To convert the format of a video.

cvb.convert_video('input.avi', 'output.mp4', vcodec='h264')

hellock/cvbase

Deprecated! Please use mmcv instead.

Introduction

Some popular features