Skip to content

Commit

Permalink
fixed paths to video embeds
Browse files Browse the repository at this point in the history
  • Loading branch information
torbjornbp committed Jul 2, 2024
1 parent 649e5d7 commit 25c1b56
Show file tree
Hide file tree
Showing 3 changed files with 7 additions and 17 deletions.
15 changes: 2 additions & 13 deletions content/post/documents/2023-08-21-text-extraction/index.md
Original file line number Diff line number Diff line change
Expand Up @@ -22,17 +22,14 @@ It's a a tool that can be used to recreate the colours of a [Kodacolor](https://
It uses a trained AI model to detect the *lenticules* on the image frames, before performing its magic.
A breath of fresh air considering all the bogus colourised black and white clips you'll see floating around these days...

---
{: data-content="KODACOLOR FILM PROCESSED BY DEEP-DOLCE"}
### Kodacolor film processed by Deep-doLCE

<video width="100%" autoplay muted loop playsinline>
<source src="kodacolor.webm" type="video/webm">
<source src="kodacolor.mp4" type="video/mp4">
<p>Your browser does not support the video format/codec.</p>
</video>

---

## GPT and Python
While I've always liked to think of myself as tech savvy, I've sadly never learned to code.
Running Python scripts has thus been an interesting learning experience.
Expand All @@ -59,17 +56,12 @@ This works remarkably well, even though you have to be patient and often rephras
It's a great learning tool, after a while I've started to get a better idea of when the bot feeds me faulty solutions (am I learning Python?).
I'm sure my involuntary tech-support friends were glad to be pestered less with Python related questions as well.

---

<video width="100%" autoplay muted loop playsinline>
<source src="demo.webm" type="video/webm">
<source src="demo.mp4" type="video/mp4">
<p>Your browser does not support the video format/codec.</p>
</video>


---

## OCR processing intertitles
This spring I had the chance to put all this to the test.
An inquiry on the AMIA-L mailing list peaked my interest.
Expand Down Expand Up @@ -117,7 +109,4 @@ Looking back at it, I'm amazed at what I managed to achieve with my limited expe

The script can be found here: [https://github.com/torbjornbp/video-ocr2srt](https://github.com/torbjornbp/video-ocr2srt)

---
{: data-content="FUNCTIONAL SRT"}

![](srtscreenshot.webp)
{{< figure src="srtscreenshot.webp" caption="Functional SRT">}}
Original file line number Diff line number Diff line change
Expand Up @@ -7,9 +7,10 @@ aliases: "/kodacolor-identification.html"
---

### Kodacolor lenticules
![Kodacolor lenticules](kodacolor_lenticules.jpg)

![](kodacolor_lenticules_zoomed.jpg)
{{< figure src="kodacolor_lenticules.jpg" alt="Kodacolor lenticules">}}

{{< figure src="kodacolor_lenticules_zoomed.jpg" alt="Kodacolor lenticules zoomed">}}

### Kodacolor sheen
<video width="100%" autoplay muted loop playsinline>
Expand All @@ -19,7 +20,7 @@ aliases: "/kodacolor-identification.html"

### Kodacolor film processed by deep-dolce
<video width="100%" autoplay muted loop playsinline>
<source src="kodacolor.webm" type="video/webm">
<source src="kodacolor.mp4" type="video/mp4">
<source src="/post/2023/08/21/video-ocr.py-and-chatgpt/kodacolor.webm" type="video/webm">
<source src="/post/2023/08/21/video-ocr.py-and-chatgpt/kodacolor.mp4" type="video/mp4">
<p>Your browser does not support the video format/codec.</p>
</video>
Binary file not shown.

0 comments on commit 25c1b56

Please sign in to comment.