denku 0.1.5

Custom computer vision utilities for image and video processing, visualization, and memory management.

DENK UTILS

Custom computer vision utilities for image and video processing, visualization, and memory management.

Installation

For Users

pip install denku

Features

Image Processing

from denku import read_image, save_image, change_contrast

# Read and process images
image = read_image("path/to/image.jpg")
processed_image = change_contrast(image, contrast=20)
save_image(processed_image, "path/to/output.jpg")

# Create and apply masks
from denku import get_color_mask_with_hsv, merge_images_by_mask_with_gauss
import numpy as np

# Create a mask based on HSV color range
color_min = np.array([0, 100, 100])
color_max = np.array([10, 255, 255])
mask = get_color_mask_with_hsv(image, color_min, color_max)

# Merge images using a mask
background = read_image("path/to/background.jpg")
merged = merge_images_by_mask_with_gauss(background, image, mask)

# Resize images
from denku import resize_to_min_side, resize_to_max_side

# Proportionally resize image to have minimum side of 256 pixels
resized_min = resize_to_min_side(image, min_side=256)

# Proportionally resize image to have maximum side of 512 pixels
resized_max = resize_to_max_side(image, max_side=512)

# Center crop and rotate images
from denku import center_crop, rotate_image, show_images

# Get original dimensions
h, w = image.shape[:2]

# Crop center 50% of the image
target_h = h // 2
target_w = w // 2
cropped = center_crop(image, target_h, target_w)

# Rotate image by different angles
rotated_90 = rotate_image(image, angle=90)    # Rotate 90 degrees counter-clockwise
rotated_45 = rotate_image(image, angle=45)    # Rotate 45 degrees counter-clockwise
rotated_neg30 = rotate_image(image, angle=-30) # Rotate 30 degrees clockwise

# Display all images in a grid
show_images(
    images=[image, cropped, rotated_90, rotated_45, rotated_neg30],
    n_rows=2,
    titles=['Original', 'Center Cropped', '90°', '45°', '-30°'],
    global_title='Image Processing Examples'
)

Video Processing

import denku

# Read video frames
# Read every frame
frames, fps = denku.read_video("video.mp4")

# Read every 2nd frame
frames, fps = denku.read_video("video.mp4", frame_stride=2)

# Read every 5th frame
frames, fps = denku.read_video("video.mp4", frame_stride=5)

# Convert video to different FPS
converted_frames = denku.convert_video_fps(frames, original_fps=fps, target_fps=16)
denku.save_video(converted_frames, "output.mp4", fps=16)

# Create a grid of videos
video_paths = ["video1.mp4", "video2.mp4", "video3.mp4", "video4.mp4"]
grid = denku.create_video_grid(video_paths, grid_size=(2, 2))
denku.save_video(grid, "video_grid.mp4")

# Advanced video operations
from denku import overlay_video, concatenate_videos, save_video_from_frames

# Overlay one video on top of another
overlay_video(
    background_path='path/to/video1.mp4',
    overlay_path='path/to/video2.mp4',
    output_path='output/overlayed.mp4',
    overlay_scale=0.3,  # Make overlay 30% of original size
    x_offset=10,        # 10 pixels from left
    y_offset=10         # 10 pixels from top
)

# Concatenate videos horizontally
concatenate_videos(
    input_paths=['path/to/video1.mp4', 'path/to/video2.mp4'],
    output_path='output/concatenated.mp4'
)

# Save frames as video
frames, fps = read_video('path/to/video.mp4')
save_video_from_frames(
    frames=frames,
    output_path='output/saved.mp4',
    fps=fps,
    codec='mp4v',
    input_format='rgb'
)

Box Operations

from denku import get_boxes_iou, get_boxes_intersection, get_boxes_union, get_boxes_iou_matrix

# Example boxes (x1, y1, x2, y2)
box1 = (10, 10, 50, 50)  # First box
box2 = (30, 30, 70, 70)  # Second box

# Calculate IoU (Intersection over Union)
iou = get_boxes_iou(box1, box2)
print(f"IoU: {iou:.2f}")  # Should be around 0.25

# Calculate intersection area
intersection = get_boxes_intersection(box1, box2)
print(f"Intersection area: {intersection}")  # Should be 400

# Calculate union area
union = get_boxes_union(box1, box2)
print(f"Union area: {union}")  # Should be 1600

# Calculate IoU matrix between two sets of boxes
boxes1 = [
    (10, 10, 50, 50),   # First box
    (20, 20, 60, 60)    # Second box
]
boxes2 = [
    (30, 30, 70, 70),   # Third box
    (40, 40, 80, 80)    # Fourth box
]

# Calculate IoU matrix between two sets of boxes
iou_matrix = get_boxes_iou_matrix(boxes1, boxes2)
print("IoU Matrix:")
print(iou_matrix)
# Output will be a 2x2 matrix showing IoU between each pair of boxes

Visualization in Jupyter Notebooks

from denku import show_image, show_images, show_video_in_jupyter, show_gif_in_jupyter

# Display images
show_image(image, title="My Image", figsize=(10, 10))

# Display multiple images in a grid
show_images([image1, image2, image3], n_rows=2, titles=["Image 1", "Image 2", "Image 3"])

# Display video in Jupyter
show_video_in_jupyter("path/to/video.mp4", width=640)

# Display GIF in Jupyter
show_gif_in_jupyter("path/to/animation.gif", width=480)

Memory Management

from denku import empty_cuda_cache, print_cuda_allocated_memory

# Reset CUDA memory and run garbage collection
empty_cuda_cache()

# Print current CUDA memory usage
print_cuda_allocated_memory()

Development

Setup Development Environment

# Clone the repository
git clone https://github.com/TheDenk/denku.git
cd denku

# Install with development dependencies
pip install -e ".[dev]"

# Install pre-commit hooks
pre-commit install

Running Tests

# Run all tests
pytest

# Run tests with coverage report
pytest --cov=denku

# Run tests for a specific module
pytest tests/test_utils.py

# Run tests with detailed output
pytest -v

# Run tests and show coverage in HTML format
pytest --cov=denku --cov-report=html

Contacts

Issues should be raised directly in the repository. For professional support and recommendations please [e-mail].