Aging/oldify a dataset of pictures Old Photo Restoration via Deep Latent Space Translation (https://paperswithcode.com/paper/old-photo-restoration-via-deep-latent-space)
My team and I are interested in reproducing their work. However, the dataset is not provided to us. We need several hundred photos and people on them. We found the following dataset IMDB-WIKI – 500k+ face images with age and gender labels.

What is the best way to oldify that type of picture?
EDIT
Here is a way to "oldify" the pictures. Using OpenCV, I can convert these images to black and white (gray scaled) and then merge/ blend these images with different old pictures textures. For that see 
Adding (blending) two images using OpenCV
. This technique seems well, but hard to apply to a large set of pictures. I would like to obtain something similar to above pictures.
 A: According to the paper, you can't.
Quoted directly from the paper:
"The
degradation process of old photos is rather complex, and
there exists no degradation model that can realistically
render the old photo artifact."
"No matter for unstructured or structured degradation,
though the above learning-based methods can achieve remarkable results, they are all trained on the synthetic data.
Therefore, their performance on the real dataset highly relies
on synthetic data quality. For real old images, since they
are often seriously degraded by a mixture of unknown
degradation, the underlying degradation process is much
more difficult to be accurately characterized. In other words,
the network trained on synthetic data only, will suffer from
the domain gap problem and perform badly on real old
photos"
"However, these
methods still rely on supervised learning from synthetic
data and hence cannot generalize to real photos."
"Old photos contain far more complex degradation that is hard to be modeled realistically and there always exists a domain gap between synthetic and real photos. As such, the network usually cannot generalize well to real photos by purely learning from synthetic data."
If this paper is to be believed, there is no known way to "oldify" a picture which will faithfully represent degradation in real photographs. A model trained on synthetic data may not apply well to real data.
A: I cannot comment on others' answers, so forgive me if it somewhat repeats what others said, but the paper does highlight a way to "oldify" the pictures, although they admit the process is impossible to recreate artificially.
Read section 4.2 (Data Generation) from the paper:

Unstructured Degradation
We  use  the  following  operations to simulate the unstructured degradation. >Specifically,

*

*Gaussian white noise withσ∈(5,50).


*Gaussian blur with kernel sizek⊂ {3,5,7}and stan-dard deviationσ∈(1.0,5.0);


*JPEG compression whose quality level in the range of(40,100);


*Color jitter which randomly shifts the RGB color chan-nels by(−20,20);


*Box blur to mimic the lens defocus.

I have tried it myself, and the results are not too bad, but they hardly apply to pictures taken with actual film.
Here's an example:
