How and why?
When working as a freelance web developer I could find the following issue with some of my clients:
- Some applications which behaviour I could not influence had issues with a picture's aspect orientation
- The images were recorded in portrait mode while being landscape in the application where they were used in. Therefore the viewer got it presented in the wrong visual orientation.
The reason for this normally lies within the EXIF metadata. These usually get created by the camera automatically.
Now, if the picture was not recorded in landscape most cameras add the correct (i.e. portrait) orientation in the EXIF orientation data. Every commonly used image editing is very well suited to correctly apply these rotation settings.
If any application, for whatever reason, removes the metadata from the image without correct interpretation the to the viewer the image seems to be rotated wrongly.
To prevent this, you can use this application to rotate the image 'physically' by the EXIF data. Which has to happen before the picture is used by the target application.
Here you can just upload (multiple) JPEG images which will be automatically rotated correctly by interpreting the EXIF data (if existing). Directly after this operation the modified image will be presented for download. All temporarily saved images will be removed directly after the operation.
This method can be called a JPEG image rotator.
More advanced rotation detection features
Recognizing if a picture has been rotated or even flipped is a way more difficult task for a computer to do.
This would include training and using trained models enabling the application to recognize if an image had been rotated, even if there were no EXIF metadata existing.
If you're interested in a feature like that, leave me a note.