How to get image height and width using java?
Solution 1:
Here is something very simple and handy.
BufferedImage bimg = ImageIO.read(new File(filename));
int width = bimg.getWidth();
int height = bimg.getHeight();
Solution 2:
This is a rewrite of the great post by @Kay, which throws IOException and provides an early exit:
/**
* Gets image dimensions for given file
* @param imgFile image file
* @return dimensions of image
* @throws IOException if the file is not a known image
*/
public static Dimension getImageDimension(File imgFile) throws IOException {
int pos = imgFile.getName().lastIndexOf(".");
if (pos == -1)
throw new IOException("No extension for file: " + imgFile.getAbsolutePath());
String suffix = imgFile.getName().substring(pos + 1);
Iterator<ImageReader> iter = ImageIO.getImageReadersBySuffix(suffix);
while(iter.hasNext()) {
ImageReader reader = iter.next();
try {
ImageInputStream stream = new FileImageInputStream(imgFile);
reader.setInput(stream);
int width = reader.getWidth(reader.getMinIndex());
int height = reader.getHeight(reader.getMinIndex());
return new Dimension(width, height);
} catch (IOException e) {
log.warn("Error reading: " + imgFile.getAbsolutePath(), e);
} finally {
reader.dispose();
}
}
throw new IOException("Not a known image file: " + imgFile.getAbsolutePath());
}
I guess my rep is not high enough for my input to be considered worthy as a reply.
Solution 3:
I tried to test performance using some of the various approaches listed. It's hard to make a rigorous test as many factors affect the result. I prepared two folders, one with 330 jpg files and another one with 330 png files. The average file size was 4Mb in both cases. Then I called getDimension for each file. Each implementation of getDimension method and each image type was tested separately (separate run). Here is the execution times that I got (first number for jpg, second number for png):
1(Apurv) - 101454ms, 84611ms
2(joinJpegs) - 471ms, N/A
3(Andrew Taylor) - 707ms, 68ms
4(Karussell, ImageIcon) - 106655ms, 100898ms
5(user350756) - 2649ms, 68ms
It's obvious that some methods load the whole file in order to get dimensions while others get by just reading some header information from the image. I think these numbers may be useful when application performance is critical.
Thank you everyone for the contribution to this thread - very helpful.
Solution 4:
I have found another way to read an image size (more generic). You can use ImageIO class in cooperation with ImageReaders. Here is the sample code:
private Dimension getImageDim(final String path) {
Dimension result = null;
String suffix = this.getFileSuffix(path);
Iterator<ImageReader> iter = ImageIO.getImageReadersBySuffix(suffix);
if (iter.hasNext()) {
ImageReader reader = iter.next();
try {
ImageInputStream stream = new FileImageInputStream(new File(path));
reader.setInput(stream);
int width = reader.getWidth(reader.getMinIndex());
int height = reader.getHeight(reader.getMinIndex());
result = new Dimension(width, height);
} catch (IOException e) {
log(e.getMessage());
} finally {
reader.dispose();
}
} else {
log("No reader found for given format: " + suffix));
}
return result;
}
Note that getFileSuffix is method that returns extension of path without "." so e.g.: png, jpg etc. Example implementation is:
private String getFileSuffix(final String path) {
String result = null;
if (path != null) {
result = "";
if (path.lastIndexOf('.') != -1) {
result = path.substring(path.lastIndexOf('.'));
if (result.startsWith(".")) {
result = result.substring(1);
}
}
}
return result;
}
This solution is very quick as only image size is read from the file and not the whole image. I tested it and there is no comparison to ImageIO.read performance. I hope someone will find this useful.
Solution 5:
You can load jpeg binary data as a file and parse the jpeg headers yourself. The one you are looking for is the 0xFFC0 or Start of Frame header:
Start of frame marker (FFC0)
* the first two bytes, the length, after the marker indicate the number of bytes, including the two length bytes, that this header contains
* P -- one byte: sample precision in bits (usually 8, for baseline JPEG)
* Y -- two bytes
* X -- two bytes
* Nf -- one byte: the number of components in the image
o 3 for color baseline JPEG images
o 1 for grayscale baseline JPEG images
* Nf times:
o Component ID -- one byte
o H and V sampling factors -- one byte: H is first four bits and V is second four bits
o Quantization table number-- one byte
The H and V sampling factors dictate the final size of the component they are associated with. For instance, the color space defaults to YCbCr and the H and V sampling factors for each component, Y, Cb, and Cr, default to 2, 1, and 1, respectively (2 for both H and V of the Y component, etc.) in the Jpeg-6a library by the Independent Jpeg Group. While this does mean that the Y component will be twice the size of the other two components--giving it a higher resolution, the lower resolution components are quartered in size during compression in order to achieve this difference. Thus, the Cb and Cr components must be quadrupled in size during decompression.
For more info about the headers check out wikipedia's jpeg entry or I got the above info here.
I used a method similar to the code below which I got from this post at the sun forums:
import java.awt.Dimension;
import java.io.*;
public class JPEGDim {
public static Dimension getJPEGDimension(File f) throws IOException {
FileInputStream fis = new FileInputStream(f);
// check for SOI marker
if (fis.read() != 255 || fis.read() != 216)
throw new RuntimeException("SOI (Start Of Image) marker 0xff 0xd8 missing");
Dimension d = null;
while (fis.read() == 255) {
int marker = fis.read();
int len = fis.read() << 8 | fis.read();
if (marker == 192) {
fis.skip(1);
int height = fis.read() << 8 | fis.read();
int width = fis.read() << 8 | fis.read();
d = new Dimension(width, height);
break;
}
fis.skip(len - 2);
}
fis.close();
return d;
}
public static void main(String[] args) throws IOException {
System.out.println(getJPEGDimension(new File(args[0])));
}
}