Byte array of a jpeg header

Question

For a particular reason I am wanting to extract the header info from a jpeg which has been loaded into a byte array.

To find the length of the header I have been comparing the bytes of different jpegs.

The common denominator is that the 1st 623 bytes of each image have the same byte value.

These jpegs are of the same height, width, resolution but of different images/scenes.

Is it really this simplistic? Are the 1st 623 the common header between all assuming that the header does not contain the length of each image in bytes?

Thanks both of you. I wanted to know, in my scenario whether I can simplistically ignore the 1st 623 bytes and the last 2 bytes and say 'the rest are the pixel info'. I have seen these header descripts before. I guess I will run a test to see if I can reconstruct the image on the server just b adding the 'common' 623 header and 'common' 2 bytes tail. — Andrew Simpson
– Andrew Simpson, Commented Dec 19, 2013 at 10:41
Your specific method might happen to work for your input data, but it's in no way a generic way to do this... ;-) JFIF supports tables-only files ("header" only), as well as table-less files ("data" only). This is probably what you want to use, but you need an encoder/decoder that supports this. — Harald K
– Harald K, Commented Dec 19, 2013 at 12:35
Hi, thanks for your feedback. I am quite sure that this would not be a generic approach. It was specific to my app that has a controlled feed of jpegs. Thanks for the info :) — Andrew Simpson
– Andrew Simpson, Commented Dec 19, 2013 at 15:59

halfer · Accepted Answer · 2018-11-17 14:40:05Z

1

There is no "JPEG" header (other than the start of image marker). JPEG is a sequence markers, some of which have following data blocks.

Various JPEG file formats define formats (e.g. JFIF, XFIF) for APPn markers that identify the file format.

Your 623 matching bytes could well be the same comment marker written by the same encoder. So, NO you cannot ignore 623 bytes and reliably determine the contents of the file.

My guess is that it would be in an APPn marker specific to the particular file format or in the Start of Frame marker.

edited Nov 17, 2018 at 14:40

halfer

20.2k20 gold badges110 silver badges207 bronze badges

answered Mar 26, 2014 at 4:13

user3344003

21.9k3 gold badges29 silver badges68 bronze badges

Sign up to request clarification or add additional context in comments.

Comments

halfer · Accepted Answer · 2018-11-17 14:38:29Z

1

(Posted answer on behalf of the question author).

From my tests If I extract the header from the very 1st image (which is 623 bytes long) and the last 2 bytes 255, 217 and then block copy the bytes in between (623 to length of byte array -2) and append the extracted (1st) header and tail on the server side the image processing works.

answered Nov 17, 2018 at 14:38

community wiki

halfer

Collectives™ on Stack Overflow

Byte array of a jpeg header

2 Answers 2

Comments

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

Comments

Comments

Your Answer

Sign up or log in

Post as a guest

Related