README.md
1# This package is no longer being updated! Please look for alternatives if that bothers you.
2
3Resize
4======
5
6Image resizing for the [Go programming language](http://golang.org) with common interpolation methods.
7
8[](https://travis-ci.org/nfnt/resize)
9
10Installation
11------------
12
13```bash
14$ go get github.com/nfnt/resize
15```
16
17It's that easy!
18
19Usage
20-----
21
22This package needs at least Go 1.1. Import package with
23
24```go
25import "github.com/nfnt/resize"
26```
27
28The resize package provides 2 functions:
29
30* `resize.Resize` creates a scaled image with new dimensions (`width`, `height`) using the interpolation function `interp`.
31 If either `width` or `height` is set to 0, it will be set to an aspect ratio preserving value.
32* `resize.Thumbnail` downscales an image preserving its aspect ratio to the maximum dimensions (`maxWidth`, `maxHeight`).
33 It will return the original image if original sizes are smaller than the provided dimensions.
34
35```go
36resize.Resize(width, height uint, img image.Image, interp resize.InterpolationFunction) image.Image
37resize.Thumbnail(maxWidth, maxHeight uint, img image.Image, interp resize.InterpolationFunction) image.Image
38```
39
40The provided interpolation functions are (from fast to slow execution time)
41
42- `NearestNeighbor`: [Nearest-neighbor interpolation](http://en.wikipedia.org/wiki/Nearest-neighbor_interpolation)
43- `Bilinear`: [Bilinear interpolation](http://en.wikipedia.org/wiki/Bilinear_interpolation)
44- `Bicubic`: [Bicubic interpolation](http://en.wikipedia.org/wiki/Bicubic_interpolation)
45- `MitchellNetravali`: [Mitchell-Netravali interpolation](http://dl.acm.org/citation.cfm?id=378514)
46- `Lanczos2`: [Lanczos resampling](http://en.wikipedia.org/wiki/Lanczos_resampling) with a=2
47- `Lanczos3`: [Lanczos resampling](http://en.wikipedia.org/wiki/Lanczos_resampling) with a=3
48
49Which of these methods gives the best results depends on your use case.
50
51Sample usage:
52
53```go
54package main
55
56import (
57 "github.com/nfnt/resize"
58 "image/jpeg"
59 "log"
60 "os"
61)
62
63func main() {
64 // open "test.jpg"
65 file, err := os.Open("test.jpg")
66 if err != nil {
67 log.Fatal(err)
68 }
69
70 // decode jpeg into image.Image
71 img, err := jpeg.Decode(file)
72 if err != nil {
73 log.Fatal(err)
74 }
75 file.Close()
76
77 // resize to width 1000 using Lanczos resampling
78 // and preserve aspect ratio
79 m := resize.Resize(1000, 0, img, resize.Lanczos3)
80
81 out, err := os.Create("test_resized.jpg")
82 if err != nil {
83 log.Fatal(err)
84 }
85 defer out.Close()
86
87 // write new image to file
88 jpeg.Encode(out, m, nil)
89}
90```
91
92Caveats
93-------
94
95* Optimized access routines are used for `image.RGBA`, `image.NRGBA`, `image.RGBA64`, `image.NRGBA64`, `image.YCbCr`, `image.Gray`, and `image.Gray16` types. All other image types are accessed in a generic way that will result in slow processing speed.
96* JPEG images are stored in `image.YCbCr`. This image format stores data in a way that will decrease processing speed. A resize may be up to 2 times slower than with `image.RGBA`.
97
98
99Downsizing Samples
100-------
101
102Downsizing is not as simple as it might look like. Images have to be filtered before they are scaled down, otherwise aliasing might occur.
103Filtering is highly subjective: Applying too much will blur the whole image, too little will make aliasing become apparent.
104Resize tries to provide sane defaults that should suffice in most cases.
105
106### Artificial sample
107
108Original image
109
110
111<table>
112<tr>
113<th><img src="http://nfnt.github.com/img/rings_300_NearestNeighbor.png" /><br>Nearest-Neighbor</th>
114<th><img src="http://nfnt.github.com/img/rings_300_Bilinear.png" /><br>Bilinear</th>
115</tr>
116<tr>
117<th><img src="http://nfnt.github.com/img/rings_300_Bicubic.png" /><br>Bicubic</th>
118<th><img src="http://nfnt.github.com/img/rings_300_MitchellNetravali.png" /><br>Mitchell-Netravali</th>
119</tr>
120<tr>
121<th><img src="http://nfnt.github.com/img/rings_300_Lanczos2.png" /><br>Lanczos2</th>
122<th><img src="http://nfnt.github.com/img/rings_300_Lanczos3.png" /><br>Lanczos3</th>
123</tr>
124</table>
125
126### Real-Life sample
127
128Original image
129
130
131<table>
132<tr>
133<th><img src="http://nfnt.github.com/img/IMG_3694_300_NearestNeighbor.png" /><br>Nearest-Neighbor</th>
134<th><img src="http://nfnt.github.com/img/IMG_3694_300_Bilinear.png" /><br>Bilinear</th>
135</tr>
136<tr>
137<th><img src="http://nfnt.github.com/img/IMG_3694_300_Bicubic.png" /><br>Bicubic</th>
138<th><img src="http://nfnt.github.com/img/IMG_3694_300_MitchellNetravali.png" /><br>Mitchell-Netravali</th>
139</tr>
140<tr>
141<th><img src="http://nfnt.github.com/img/IMG_3694_300_Lanczos2.png" /><br>Lanczos2</th>
142<th><img src="http://nfnt.github.com/img/IMG_3694_300_Lanczos3.png" /><br>Lanczos3</th>
143</tr>
144</table>
145
146
147License
148-------
149
150Copyright (c) 2012 Jan Schlicht <janschlicht@gmail.com>
151Resize is released under a MIT style license.
152
){kind=link}
){kind=link}
){kind=link}
