cvvideoplayer 1.1.4

moduler multi purpose video player for debugging algorithms in python

Introduction

CV video player is a Python-based customizable video player that helps computer vision practitioners to develop, analyze, and debug their video-related algorithms and models.

The video player is interactive, operating only with keyboard presses (no UI buttons). The user can register shortcuts using the VideoPlayer class's API.

The player is designed as a callback system. When initialized the player receives a list of callbacks, each with optional hooks for setup, before frame resize, and after frame resize. on runtime the callbacks hooks will run in order as specified in the input list.

See it in action

Running frame by frame

Play/Pause and control play speed and direction

Draw bounding boxes and adjust labels

Installation

pip install cvvideoplayer

Quick Start

from cvvideoplayer import create_video_player
from cvvideoplayer.frame_editors import FrameInfoOverlay, HistogramEqualizer, FrameNormalizer, KeyMapOverlay

VIDEO_OR_FRAME_FOLDER_PATH = "<add local path here>"

def run_player():
    video_player = create_video_player(
        video_source=VIDEO_OR_FRAME_FOLDER_PATH,
        frame_edit_callbacks=[
            FrameInfoOverlay(),
            HistogramEqualizer(),
            FrameNormalizer(),
            KeyMapOverlay(),
        ],
        record=True,
    )

    video_player.run()


if __name__ == "__main__":
    run_player()

In this example, we initiate a very basic video player that will play "example_video.mp4" with added basic frame edit callbacks:

• FrameInfoOverlay: Prints the current frame number and original frame resolution in the top left corner
• HistogramEqualizer: Preforms histogram equalization on the image
• FrameNormalizer: Allows the user to set a dynamic range for the image.
• KeyMapOverlay: prints all optional shortcuts registered by all callbacks

Check out the ./demos folder which shows the use of other cool frame edit callback such as OpticalFlow and DetectionCsvPlotter