14 ms·
You can also generate similar scale astronomical images with GANs! I worked on a model that could do this last year: Paper: https://arxiv.org/abs/1904.10286 h
by Smith42 6y ago
You can also generate similar scale astronomical images with GANs!
I worked on a model that could do this last year:
Paper: https://arxiv.org/abs/1904.10286 https://arxiv.org/abs/1904.10286
7.6 gigapixel image: https://star.herts.ac.uk/~jgeach/gdf https://star.herts.ac.uk/~jgeach/gdf
- mncharity 6y ago> RGB channels corresponding to the F814W, F775W and F606W bands. [from paper] IR, red, and orange. In case anyone else zoomed in, saw odd colors, and wondered. Anyone have suggestions for an HDR or zoomable synthetic field that's visually realistic? Last week I was exploring a try-to-stay-motivated covid project of an "if they're lots of stars, why can't I see them?" educational web app. Basically a horizon+sky slice with sliders for ambient light level (day to night to beyond inner solar system dust scatter) and for exposure (pupil narrow to dilated to dark adapted to hubble at pluto), and presets. Synthetic permits optimizing for pedagogy, and avoids issues around selecting some real region. Might be useful when teaching light pollution. I'd like it for an exploratory atoms-up early-primary learning progression variant emphasizing nuclei, thus nucleosynthesis, and a "I've seen skies of stars in video, but stars are almost absent from my lived experience" challenge. Collaboration welcome. :)