Video Shot Detection

Video shot detection is a computer vision technique used to automatically identify the boundaries between shots in a video sequence. Shot detection algorithms analyze the visual and/or auditory features of a video to determine when a new shot begins and ends, which is useful for tasks such as video indexing, summarization, and editing.

Rules

• Programs must be based on C/C++ or Python w/wo window interface

Requirements

Please complete the following two methods

1. Using histogram feature to implement the video shot detection.
2. Design your own algorithm to implement the video shot detection.
   Note : Please indicate the frame index where the shot changes.

Usage

For running original.py, adaptivethreshold.py, remember to change

• video_source - your video file path
• output_dir - your output folder path

For running histogram.py, remember to change

• input_dir - your .csv folder path

Installation

Download from github

git clone https://github.com/nick8592/Video-Shot-Detection.git

Install reqired dependencies

cd main/
pip install -r requirements.txt

Example

Run Original

python original.py

Run AdaptiveThreshold

python adaptivethreshold.py

Plot Shot Detection Histogram

python histogram.py

Flowchart

Shot Detection Histogram

KungFu Hustle

Spider Man