DynamicCropper

DynamicCropper is a Python library that allows you to dynamically crop a person in a video based on various parameters.

TODOs

This section outlines the tasks and improvements planned for the DynamicCropper project.

• Improve performance for large video files.
• Add threading to operations.
• Create verbose control.
• Better outputs and error prints.
• Logger.
• Better file structure.
• Add better requirements.txt

How to Use

Step 1: Clone the Repository

Clone the repository to your local machine using the following command:

git clone https://github.com/AteltaAI/DynamicCropper.git

Step 2: Create a Test File

Inside the repository folder, create a new Python test file (e.g., test_cropper.py). You can use the following example code in your test file:

fromdynamic_cropperimportDynamicCroppercropper=DynamicCropper(
frame_interval=10,
size_aware=True,
only_object=True
)
# 'input_path' can be a youtube link or local video file path.result=cropper.crop(
input_path="https://youtu.be/GcZJMiHds3U",
output_folder_path="output_folder"
)

Step 3: Run the Test File

Run the test file to see how DynamicCropper works.

python test_cropper.py

Showcase of Results

Original Clip

input-video.2.mp4

Test Video 1:

Code:

fromdynamic_cropperimportDynamicCroppercropper=DynamicCropper(
frame_interval=10,
size_aware=False,
only_object=False
)
result=cropper.crop(
input_path=rf"input-video (2).mp4",
output_folder_path="output_folder"
)
print(result)

Output:

output_audio_false_false.mp4

Test Video 2:

Code:

fromdynamic_cropperimportDynamicCroppercropper=DynamicCropper(
frame_interval=10,
size_aware=True,
only_object=True
)
result=cropper.crop(
input_path=rf"input-video (2).mp4",
output_folder_path="output_folder"
)
print(result)

Output:

output_audio_true_true.mp4

Test Video 3:

Code:

fromdynamic_cropperimportDynamicCroppercropper=DynamicCropper(
frame_interval=10,
size_aware=True,
only_object=False
)
result=cropper.crop(
input_path=rf"input-video (2).mp4",
output_folder_path="output_folder"
)
print(result)

Output:

output_audio_trrue_false.mp4