Welcome to Stitcher
Stitcher is a high-performance C library for multiband image merging using the Laplacian blending technique. It enables seamless blending of overlapping images, making it ideal for applications such as panorama generation, image stitching, and exposure fusion.Key Features
Multiband Blending
Advanced Laplacian pyramid blending with configurable frequency bands for seamless image merging
Feather Blending
Smooth transition blending mode for quick and efficient image composition
SIMD Optimization
High-performance operations optimized with SIMD instructions using the SIMDe library
Multi-Threading
Parallel image processing with pthread support for maximum performance
JPEG Support
Fast JPEG compression and decompression powered by libjpeg-turbo
Distance Transform
Mask refinement with distance transform for improved blending quality
What is Multiband Blending?
Multiband blending is a sophisticated image composition technique that operates on different frequency bands of an image. By decomposing images into Laplacian pyramids, Stitcher can blend high-frequency details and low-frequency colors separately, resulting in seamless transitions without visible seams or artifacts. This approach is particularly effective for:- Panorama stitching: Merge multiple photos into wide-angle panoramas
- Exposure fusion: Combine bracketed exposures into HDR-like images
- Image composition: Seamlessly blend elements from different images
Performance
Stitcher is built for speed:- SIMD-optimized operations using platform-specific instructions (AVX2, NEON)
- Multi-threaded processing for parallel execution
- Efficient memory management with minimal allocations
- Hardware-accelerated JPEG encoding/decoding via libjpeg-turbo
Platform Support
Stitcher supports multiple platforms:- macOS: x86_64 and ARM64 (Apple Silicon)
- iOS: ARM64 devices and simulators
- Android: arm64-v8a, armeabi-v7a, x86, x86_64
Next Steps
Installation
Build and install Stitcher on your platform
Quick Start
Create your first blended image in minutes