imfinfo doesnt work in R2022b, it's bugged

Hello! I am using imfinfo to matlab version R2022b and i get this error:

Failed to create file modification time character vector.

The same function works perfectly in version R2019a.

How can i fix it in version R2022b?

My command is:

x = imfinfo('my_file.tif');

Thanks in Advance

[Jan]

Does it work with other files?

This is the public forum mainly supported by voluntary Matlab users. We cannot fix bugs of Matlab's toolbox functions. So please contact the technical support of MathWorks. You find a "Contact Us" link on the bottom of this page.

[IOANNIS]

Thanks for the quick response. Yes it works with other files.

However for this file for example https://we.tl/t-IIQ1adldGI

it doesnt work in R2022b, but it works in R2019a. I cant understand...

[Stephen23]

"However for this file for example"

Many users here will not download from dodgy third-party websites. You should upload the file here.

[IOANNIS]

File:my_file.tif size exceeds 5 MB.

[Jan]

As said already: Please send this as bug report to MathWorks.

Answer 1

[Steven Lord]

Since you said imfinfo works for other files but not this one, I'm wondering if the file has somehow been corrupted or if it was written in a non-standard way that imfinfo was lenient enough to allow in the past but is strict enough now (through bug fixes or updates) that it no longer accepts the non-standard file.

First let's check to see if this problem is specific to the .tif file processing code. MATLAB includes one .tif file as part of one of the demos, example.tif. Let's see if imfinfo works on that file. What do you see when you run these lines of code?

name = fullfile(matlabroot, 'toolbox', 'matlab', 'demos', 'example.tif');
S = imfinfo(name);
S(1).FileModDate

ans = '08-Nov-2009 01:02:06'

If that works, then let's see if some other software recognizes this file. What operating system are you using? If you're using Windows, right-click on your my_file.tif in File Explorer and select Properties. In the Properties window for your file, select the Details tab and scroll down to the end of the list, the section with the heading File. What do you see for the "Date created" and "Date modified" entries?

If you see the correct or reasonable data in the Windows properties for the file, let's try one more step. Call imfinfo on your file and let it error. Then run this command and show us what it displays. This will let us see the stack of the error and check whether you've overloaded (or downloaded an overload for) a function upon which imfinfo depends that's preventing it from operating correctly.

ex = MException.last;
result = struct2table(ex.stack)

[IOANNIS]

Hi, thanks for the response and the suggestions

I use windows 10. Everything seems fine in QGIS and in Windows Properties.

Exeption in R2022b is:

indetinfier: imageio:tifftagsread:strftimeFailure

Message: Failed to create file modification time character vector.

Using imfinfo in R2019a i noticed this line:

FileModDate: '17-????-2022 07:21:59'

So i guess the questionmarks is the problem here.

Anyway, i dont need this info from imfinfo, is it possible to skip it somehow and make it work like in R2019a?

thanks in advance

[DGM]

Can you update the modification time to something valid?

Idk if windows has a utility like touch or not.

[IOANNIS]

No i can't. it's an output of a software that is not under my control. TB of data in total with the same issue. I guess i will stay in R2019a version if its not possible in R2022b.

[Stephen23]

The image file seems to be corrupted in some way, I cannot tell exactly how. For example, when opening using GIMP it shows many data import errors:

I will check now using some EXIF tools, perhaps it lies in that part of the file.

[IOANNIS]

iminfo in R2019a:

FileModDate: '17-????-2022 07:21:59'

FileSize: 9658530

Format: 'tif'

FormatVersion: []

Width: 1098

Height: 1098

BitDepth: 64

ColorType: 'grayscale'

FormatSignature: [73 73 42 0]

ByteOrder: 'little-endian'

NewSubFileType: 0

BitsPerSample: [32 32]

Compression: 'Uncompressed'

PhotometricInterpretation: 'BlackIsZero'

StripOffsets: [1×1098 double]

SamplesPerPixel: 2

RowsPerStrip: 1

StripByteCounts: [1×1098 double]

XResolution: []

YResolution: []

ResolutionUnit: 'Inch'

Colormap: []

PlanarConfiguration: 'Chunky'

TileWidth: []

TileLength: []

TileOffsets: []

TileByteCounts: []

Orientation: 1

FillOrder: 1

GrayResponseUnit: 0.0100

MaxSampleValue: [4.2950e+09 4.2950e+09]

MinSampleValue: [0 0]

Thresholding: 1

Offset: 9651578

ExtraSamples: 0

SampleFormat: {'IEEE floating point' 'IEEE floating point'}

ModelPixelScaleTag: [100 100 0]

ModelTiepointTag: [0 0 0 300000 700020 0]

GeoKeyDirectoryTag: [1 1 0 7 1024 0 1 1 1025 0 1 1 1026 34737 22 0 2049 34737 7 22 2054 0 1 9102 3072 0 1 32630 3076 0 1 9001]

GeoAsciiParamsTag: 'WGS 84 / UTM zone 30N|WGS 84|'

In QGIS everything is fine also.

The error in R2022b is comming from these lines in imfinfo:

try

info = feval(fmt_s.info, filename);

catch ME

if isUrl

msgtext = replace(ME.message, filename, source);

newME = MException(ME.identifier, msgtext);

else

newME = ME;

end

throwAsCaller(newME); <------ here, this throw error in R2022, but not in R2019

end

[Jan]

This means, that "FileModDate: '17-????-2022 07:21:59'" was accepted in R2019b, but not in R2022b. I would not consider this as a bug, but the tool, which have created this file is buggy.

[Steven Lord]

I suspect that release R2022b throwing an error for this file with invalid information may be a result of a bug fix in that release for a crash introduced in release R2022a. See Bug Report 2724274. This is the only Bug Report I found searching the past 5 years for "imfinfo".

If that other software is writing TIF files with errors in the data and/or metadata, as seems to be the case based on the file modification information you showed and Stephen23's experiment with GIMP, you probably should report this to the authors of that software. Since multiple software packages (MATLAB and GIMP) reported problems with this specific file I don't think this is a bug in MATLAB.

[Stephen23]

I tried some command-line and Python-based tools, they all complain about exactly the same things:

I am not an expert on image file formats, but clearly there is something wrong with this one.

[DGM]

I suppose it depends which tags you need to read.

Does geotiffinfo() rely on the same underlying resources as imfinfo()? I have no way of testing this from R2019b, but considering what this file looks like, that might be something to look at.

If nothing else, you might be able to use system('exiftool myfile.tif') or something to get some info as a workaround.

[IOANNIS]

Thank you @DGM !!! geotiffinfo() works perfect in R2022b for these files. I had completely forgotten about this function. Thank you!