This Python script processes a video by removing frames that match a specified reference image and then compresses the resulting video. The script supports GPU acceleration for compression if an NVIDIA or AMD GPU is available, falling back to CPU-based compression otherwise.
version with opencv mathcing operation is x3 slower than cpu version. i dunno what's up with that.
- Frame Removal: Removes frames from the video that match a reference image using template matching.
- GPU Acceleration: Utilizes GPU acceleration for the final video compression task if an appropriate GPU is detected (CUDA).
- Batch Processing: Handles video frames in batches to optimize performance.
- Install Required Python Libraries:
Ensure you have Python 3.x installed and then install the required libraries using
pip3:pip install opencv-python numpy
- Install FFMPEG:
Make sure ffmpeg is installed on your system. You can download it from FFMPEG's official website or install it using a package manager.
-
Place Your Files:
Save the reference image as reference.png in the same directory as the script. Place your input video file as input.mp4 in the same directory.
-
Run the Script:
Execute the script using Python:
python video_frame_removal_[noCUDA_on_opencv].py
-
Output Files:
Intermediate video: temp_video.mp4 (saved in the temp_files directory) Final processed video: input_trimmed.mp4
Setup and Reference Image Loading: Creates a directory for temporary files if it doesn't exist. Loads and resizes the reference image to speed up processing.
Frame Processing: Each frame of the video is compared to the reference image. Frames that match the reference image are removed. Non-matching frames are written to a temporary video file in batches.
Video Compression: Detects available GPU hardware (NVIDIA or AMD) to use hardware-accelerated video encoding. Uses ffmpeg to compress the temporary video file into the final output video. If no GPU is available, falls back to CPU-based compression.
Resource Cleanup: Intermediate files and resources are cleaned up after processing.
Reference Image Resize Factor: Adjust resize_factor to change the size of the reference image for template matching. A lower factor speeds up processing but may reduce accuracy.
Compression Settings: Modify the -b:v option in the ffmpeg_command to change the video bitrate. Adjust -r to set the desired frame rate.
GPU Detection: The detect_gpu function identifies the available GPU and selects the appropriate codec. You can modify this function to add support for other GPUs or codecs.
Reference Image Not Found: Ensure reference.png is in the same directory as the script and is a valid image file.
FFMPEG Errors: Verify that ffmpeg is installed and accessible from your system's PATH.
This script is provided "as-is" without any warranties. Use at your own risk.