In the last lesson, we looked at how open-source machine learning models in Nuke can assist with ๐—ฏ๐—ฎ๐˜€๐—ถ๐—ฐ ๐—ต๐˜‚๐—บ๐—ฎ๐—ป ๐—บ๐—ฎ๐˜๐˜๐—ถ๐—ป๐—ด. Now, weโ€™re taking that same footage into Boris FX ๐— ๐—ผ๐—ฐ๐—ต๐—ฎ ๐—ฃ๐—ฟ๐—ผ ๐Ÿฎ๐Ÿฌ๐Ÿฎ๐Ÿฑ to test its new ๐— ๐—ฎ๐˜€๐—ธ ๐— ๐—Ÿ feature.

Hereโ€™s what it does:

Instead of manually creating a mask, ๐˜†๐—ผ๐˜‚ ๐—ฎ๐—ฑ๐—ฑ ๐— ๐—ฎ๐˜€๐—ธ ๐— ๐—Ÿ ๐˜๐—ผ ๐—ฎ ๐—น๐—ฎ๐˜†๐—ฒ๐—ฟ, ๐—ฎ๐—ป๐—ฑ ๐—ถ๐˜ ๐—ด๐—ฒ๐—ป๐—ฒ๐—ฟ๐—ฎ๐˜๐—ฒ๐˜€ ๐—ฎ ๐—บ๐—ฎ๐—ฐ๐—ต๐—ถ๐—ป๐—ฒ-๐—ฎ๐˜€๐˜€๐—ถ๐˜€๐˜๐—ฒ๐—ฑ ๐—บ๐—ฎ๐˜๐˜๐—ฒ ๐—ฎ๐˜‚๐˜๐—ผ๐—บ๐—ฎ๐˜๐—ถ๐—ฐ๐—ฎ๐—น๐—น๐˜†. No need to track points, no need to animate splines frame by frame, just let the tool do its job.

But letโ€™s be clear: ๐˜๐—ต๐—ถ๐˜€ ๐—ถ๐˜€ ๐—ป๐—ผ๐˜ ๐—ฎ ๐—บ๐—ฎ๐—ด๐—ถ๐—ฐ ๐—ฏ๐˜‚๐˜๐˜๐—ผ๐—ป.

Since this is a matting tool, ๐˜๐—ต๐—ฒ ๐—ฎ๐—น๐—ฝ๐—ต๐—ฎ ๐˜†๐—ผ๐˜‚ ๐—ด๐—ฒ๐˜ ๐—ถ๐˜€ ๐—บ๐—ผ๐˜€๐˜๐—น๐˜† ๐—ฎ ๐˜€๐—ผ๐—น๐—ถ๐—ฑ ๐˜€๐—ต๐—ฎ๐—ฝ๐—ฒ, ๐—ฑ๐—ฒ๐˜ƒ๐—ผ๐—ถ๐—ฑ ๐—ผ๐—ณ ๐—บ๐—ผ๐˜๐—ถ๐—ผ๐—ป ๐—ฏ๐—น๐˜‚๐—ฟ ๐˜๐—ฟ๐—ฎ๐—ถ๐—น๐˜€. Itโ€™s not replacing fine detail work yet, but itโ€™s changing how we ๐—ฎ๐—ฝ๐—ฝ๐—ฟ๐—ผ๐—ฎ๐—ฐ๐—ต ๐˜๐—ต๐—ฒ ๐—ฝ๐—ฟ๐—ผ๐—ฐ๐—ฒ๐˜€๐˜€.

And thatโ€™s the point.

These are ๐—š๐—ฒ๐—ป ๐Ÿญ ๐˜๐—ผ๐—ผ๐—น๐˜€. Right now, theyโ€™re assisting. But I can already see a future where we donโ€™t just use generic training data, we ๐˜๐—ฟ๐—ฎ๐—ถ๐—ป ๐—ผ๐˜‚๐—ฟ ๐—ผ๐˜„๐—ป ๐—บ๐—ผ๐—ฑ๐—ฒ๐—น๐˜€, fine-tuned to our workflow. Where we donโ€™t just pull from local sources, we ๐—ฝ๐˜‚๐—น๐—น ๐—ณ๐—ฟ๐—ผ๐—บ ๐˜๐—ต๐—ฒ ๐—ฐ๐—น๐—ผ๐˜‚๐—ฑ, accessing a library of pre-trained AI masks that improve over time.

This is the beginning of ๐—ฉ๐—™๐—ซ ๐˜€๐—ต๐—ถ๐—ณ๐˜๐—ถ๐—ป๐—ด ๐—ฎ๐˜„๐—ฎ๐˜† ๐—ณ๐—ฟ๐—ผ๐—บ ๐—ฏ๐—ฟ๐˜‚๐˜๐—ฒ ๐—ณ๐—ผ๐—ฟ๐—ฐ๐—ฒ ๐—บ๐—ฎ๐—ป๐˜‚๐—ฎ๐—น ๐˜„๐—ผ๐—ฟ๐—ธ toward ๐—บ๐—ฎ๐—ฐ๐—ต๐—ถ๐—ป๐—ฒ-๐—ฎ๐˜€๐˜€๐—ถ๐˜€๐˜๐—ฒ๐—ฑ ๐—ฎ๐—ฟ๐˜๐—ถ๐˜€๐˜๐—ฟ๐˜†.

And thatโ€™s the real impact.

Already, these tools are allowing me to ๐˜€๐—ฝ๐—ฒ๐—ป๐—ฑ ๐—บ๐—ผ๐—ฟ๐—ฒ ๐˜๐—ถ๐—บ๐—ฒ ๐—บ๐—ฎ๐—ธ๐—ถ๐—ป๐—ด ๐—ฐ๐—ฟ๐—ฒ๐—ฎ๐˜๐—ถ๐˜ƒ๐—ฒ ๐—ฑ๐—ฒ๐—ฐ๐—ถ๐˜€๐—ถ๐—ผ๐—ป๐˜€ ๐—ถ๐—ป๐˜€๐˜๐—ฒ๐—ฎ๐—ฑ ๐—ผ๐—ณ ๐—ด๐—ฒ๐˜๐˜๐—ถ๐—ป๐—ด ๐˜€๐˜๐˜‚๐—ฐ๐—ธ ๐—ผ๐—ป ๐˜€๐—ฒ๐—ฎ๐—บ๐—น๐—ฒ๐˜€๐˜€ ๐—ฒ๐˜…๐˜๐—ฟ๐—ฎ๐—ฐ๐˜๐—ถ๐—ผ๐—ป.

It feels like ๐˜๐—ฟ๐—ฎ๐˜ƒ๐—ฒ๐—น๐—ถ๐—ป๐—ด ๐˜๐—ผ ๐—ฎ๐—ป๐—ผ๐˜๐—ต๐—ฒ๐—ฟ ๐˜€๐˜๐—ฎ๐˜๐—ฒ ๐—ผ๐—ฟ ๐—ฐ๐—ผ๐˜‚๐—ป๐˜๐—ฟ๐˜†.

You ๐—ฐ๐—ฎ๐—ป drive yourself there.
Or you can ๐˜๐—ฎ๐—ธ๐—ฒ ๐—ฎ ๐—ฝ๐—น๐—ฎ๐—ป๐—ฒ ๐—ฎ๐—ป๐—ฑ ๐—ฎ ๐—ฐ๐—ฎ๐—ฏ and get straight to where you need to be.

Thatโ€™s what these tools are doing, theyโ€™re ๐—ฐ๐—ต๐—ฎ๐—ป๐—ด๐—ถ๐—ป๐—ด ๐—ต๐—ผ๐˜„ ๐˜„๐—ฒ ๐—ด๐—ฒ๐˜ ๐˜๐—ต๐—ฒ๐—ฟ๐—ฒ.

This is where ๐— ๐—”๐—ซ (๐— ๐—ฎ๐—ฐ๐—ต๐—ถ๐—ป๐—ฒ ๐—”๐˜€๐˜€๐—ถ๐˜€๐˜๐—ฒ๐—ฑ ๐—ฉ๐—™๐—ซ) continues. Letโ€™s get into it.

Watch it on YouTube (Higher Res) - https://lnkd.in/gaqAvt6g