purecv 0.1.0

A pure Rust, high-performance computer vision library focused on safety and portability.
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102

# PureCV


![PureCv Banner](./assets/purecv_banner.png)

[![Rust CI](https://github.com/webarkit/purecv/actions/workflows/rust.yml/badge.svg)](https://github.com/webarkit/purecv/actions/workflows/rust.yml)

A high-performance, **pure Rust** computer vision library focusing on the `core` and `imgproc` modules of OpenCV. **PureCV** is built from the ground up to be memory-safe, thread-safe, and highly portable without the overhead of C++ FFI.

## 🎯 Philosophy


Unlike existing wrappers, **PureCV** is a native rewrite. It aims to provide:

* **Zero-FFI:** No complex linking or C++ toolchain requirements.
* **Memory Safety:** Elimination of segmentation faults and buffer overflows via Rust's ownership model.
* **Modern Parallelism:** Native integration with **Rayon** for effortless multi-core processing.
* **WASM & Native SIMD:** Optimized for the web and high-performance native architectures using explicit SIMD (via `pulp`).

## ✨ Features


### `purecv-core`

- **Matrix Operations:** Multi-dimensional `Matrix<T>` with support for common arithmetic (`add`, `sub`, `mul`, `div`) and bitwise logic.
- **Factory Methods:** Intuitive initialization with `zeros`, `ones`, `eye`, and `diag`.
- **Structural:** `flip`, `rotate`, `transpose`, `repeat`, `reshape`, `hconcat`, `vconcat`.
- **Math & Stats:** `sqrt`, `exp`, `log`, `pow`, `sum`, `mean`, `minMaxLoc`, `norm`.
- **Channel Management:** `split`, `merge`, `mixChannels`.

### `purecv-imgproc`

- **Color Conversions:** High-performance `cvt_color` supporting RGB, BGR, Gray, and more.
- **Filtering:** (In Progress) Fast convolutions and image filters.

## 🚀 Getting Started


### Installation


Add the following to your `Cargo.toml`:

```toml
[dependencies]
purecv = { git = "https://github.com/webarkit/purecv" }
```

### Usage Example


```rust
use purecv::core::{Matrix, Size, Scalar};
use purecv::imgproc::{cvt_color, ColorConversionCodes};

fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Create a 3-channel matrix initialized to ones
    let mat = Matrix::<f32>::ones(480, 640, 3);
    
    // Create an identity matrix
    let identity = Matrix::<f32>::eye(3, 3, 1);
    
    println!("Matrix size: {}x{}", mat.cols, mat.rows);
    Ok(())
}
```

### Running Examples


Explore the capabilities of PureCV by running the provided examples:

```bash
# Basic matrix arithmetic

cargo run --example arithmetic

# Structural operations (flip, rotate, split/merge)

cargo run --example structural_ops

# Color conversion (RGB to Grayscale)

cargo run --example color_conversion
```

## 🧪 Testing & Benchmarking


### Running Tests

PureCV uses a comprehensive suite of unit tests to ensure correctness and parity with OpenCV.

```bash
# Run all tests

cargo test
```

### Running Benchmarks

Performance is a core focus. You can run benchmarks to see the impact of SIMD and multi-threading on your architecture:

```bash
# Run all benchmarks

cargo bench
```

## 🗺 Roadmap


- [x] **Phase 1: Core Foundation** - Matrix types, arithmetic, geometric utilities, and basic structural transforms.
- [/] **Phase 2: Performance** - SIMD optimizations and benchmarking vs OpenCV C++.
- [ ] **Phase 3: WebAssembly** - Specialized wrappers and multi-threading for the web.
- [ ] **Phase 4: Image Processing** - Advanced filtering, convolutions, and feature detection.

## 📄 License


This project is licensed under the LGPL-3.0 License.