You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@tika.apache.org by Tyler Salwierz <th...@gmail.com> on 2023/10/29 13:16:27 UTC
Tika not generating OCR on HEIC images
I’m using fscrawler which uses Tika and it’s not generating OCR on heic images. The actual image metadata is indexed but the content is empty.
Is there any fix for this if it is a Tika bug?
https://pastebin.com/raw/Jp5kBi5M
Re: Tika not generating OCR on HEIC images
Posted by Luís Filipe Nassif <lf...@gmail.com>.
Sorry, actually webp is supported by tesseract directly using libwebp
(which caused a recent RCE vulnerability in a lot of tools, including
tesseract itself...)
Em sáb, 4 de nov de 2023 00:23, Luís Filipe Nassif <lf...@gmail.com>
escreveu:
> A possible future improvement is to use ImageMagick, if available on PATH,
> to convert formats not supported by java to PNG before OCR. We do this for
> heic, psd, svg, emf, wmf, webp (also supported by twelvemonkeys) and a few
> others.
>
> Luis
>
> Em dom, 29 de out de 2023 17:40, Tilman Hausherr <TH...@t-online.de>
> escreveu:
>
>> On 29.10.2023 18:00, Tyler Salwierz wrote:
>>
>> Is Apple using their own custom ocr scanner then because spotlight does
>> ocr locally on Heic.
>>
>> if they can display HEIC, then they can also convert it.
>>
>> They have an OCR:
>>
>>
>> https://developer.apple.com/documentation/vision/recognizing_text_in_images
>> Tilman
>>
>>
>>
>> On Oct 29, 2023, at 9:12 AM, Tilman Hausherr <TH...@t-online.de>
>> <TH...@t-online.de> wrote:
>>
>>
>> On 29.10.2023 14:16, Tyler Salwierz wrote:
>>
>> I’m using fscrawler which uses Tika and it’s not generating OCR on heic
>> images. The actual image metadata is indexed but the content is empty.
>>
>> Is there any fix for this if it is a Tika bug?
>>
>> https://pastebin.com/raw/Jp5kBi5M
>>
>> Heic isn't supported by tesseract, thus it isn't a bug.
>>
>> https://github.com/tesseract-ocr/tesseract/issues/2930
>>
>> https://tesseract-ocr.github.io/tessdoc/InputFormats.html
>>
>> Tilman
>>
>>
>>
>>
Re: Tika not generating OCR on HEIC images
Posted by Luís Filipe Nassif <lf...@gmail.com>.
A possible future improvement is to use ImageMagick, if available on PATH,
to convert formats not supported by java to PNG before OCR. We do this for
heic, psd, svg, emf, wmf, webp (also supported by twelvemonkeys) and a few
others.
Luis
Em dom, 29 de out de 2023 17:40, Tilman Hausherr <TH...@t-online.de>
escreveu:
> On 29.10.2023 18:00, Tyler Salwierz wrote:
>
> Is Apple using their own custom ocr scanner then because spotlight does
> ocr locally on Heic.
>
> if they can display HEIC, then they can also convert it.
>
> They have an OCR:
>
> https://developer.apple.com/documentation/vision/recognizing_text_in_images
> Tilman
>
>
>
> On Oct 29, 2023, at 9:12 AM, Tilman Hausherr <TH...@t-online.de>
> <TH...@t-online.de> wrote:
>
>
> On 29.10.2023 14:16, Tyler Salwierz wrote:
>
> I’m using fscrawler which uses Tika and it’s not generating OCR on heic
> images. The actual image metadata is indexed but the content is empty.
>
> Is there any fix for this if it is a Tika bug?
>
> https://pastebin.com/raw/Jp5kBi5M
>
> Heic isn't supported by tesseract, thus it isn't a bug.
>
> https://github.com/tesseract-ocr/tesseract/issues/2930
>
> https://tesseract-ocr.github.io/tessdoc/InputFormats.html
>
> Tilman
>
>
>
>
Re: Tika not generating OCR on HEIC images
Posted by Tilman Hausherr <TH...@t-online.de>.
On 29.10.2023 18:00, Tyler Salwierz wrote:
> Is Apple using their own custom ocr scanner then because spotlight
> does ocr locally on Heic.
if they can display HEIC, then they can also convert it.
They have an OCR:
https://developer.apple.com/documentation/vision/recognizing_text_in_images
Tilman
>
>> On Oct 29, 2023, at 9:12 AM, Tilman Hausherr <TH...@t-online.de>
>> wrote:
>>
>>
>> On 29.10.2023 14:16, Tyler Salwierz wrote:
>>> I’m using fscrawler which uses Tika and it’s not generating OCR on
>>> heic images. The actual image metadata is indexed but the content is
>>> empty.
>>>
>>> Is there any fix for this if it is a Tika bug?
>>>
>>> https://pastebin.com/raw/Jp5kBi5M
>>
>> Heic isn't supported by tesseract, thus it isn't a bug.
>>
>> https://github.com/tesseract-ocr/tesseract/issues/2930
>>
>> https://tesseract-ocr.github.io/tessdoc/InputFormats.html
>>
>> Tilman
>>
>>
Re: Tika not generating OCR on HEIC images
Posted by Tyler Salwierz <th...@gmail.com>.
Is Apple using their own custom ocr scanner then because spotlight does ocr locally on Heic.
> On Oct 29, 2023, at 9:12 AM, Tilman Hausherr <TH...@t-online.de> wrote:
>
>
> On 29.10.2023 14:16, Tyler Salwierz wrote:
>> I’m using fscrawler which uses Tika and it’s not generating OCR on heic images. The actual image metadata is indexed but the content is empty.
>>
>> Is there any fix for this if it is a Tika bug?
>>
>> https://pastebin.com/raw/Jp5kBi5M
> Heic isn't supported by tesseract, thus it isn't a bug.
>
> https://github.com/tesseract-ocr/tesseract/issues/2930
>
> https://tesseract-ocr.github.io/tessdoc/InputFormats.html
>
> Tilman
>
>
Re: Tika not generating OCR on HEIC images
Posted by Tilman Hausherr <TH...@t-online.de>.
On 29.10.2023 14:16, Tyler Salwierz wrote:
> I’m using fscrawler which uses Tika and it’s not generating OCR on
> heic images. The actual image metadata is indexed but the content is
> empty.
>
> Is there any fix for this if it is a Tika bug?
>
> https://pastebin.com/raw/Jp5kBi5M
Heic isn't supported by tesseract, thus it isn't a bug.
https://github.com/tesseract-ocr/tesseract/issues/2930
https://tesseract-ocr.github.io/tessdoc/InputFormats.html
Tilman