Moved out of the intro (`Draft — 01-01`) on 2026-06-25: the one-axis placement of every part on the classical → learned → generative spectrum was the author's rough qualitative estimate, not a measurement, and needs rethinking (per-part data, a cleaner visual, or folding it into the Machine-learning / Deep-learning chapters) before it returns to the book. Figure id: `fig-learning-transition`.

in multiple images?

Something about plugins

Something about software engineering and image processing libraries

Zero lag shutter

Beautification????

Decluttering

Hybrid images

Depth map extraction, especially for fake Depth of field

Moved here from the former *Advanced computational photography* part — dehazing, rain/snow removal, and through-the-weather imaging still need a home chapter (fold into edge-aware / single-image restoration, or give its own treatment).

---

• **how**: silicon photodiodes respond well into the **near-infrared**, so every camera ships with an **IR-cut filter** to keep IR from contaminating color. Two routes to NIR imaging: (1) **convert the camera** — remove the internal IR-cut filter (and optionally replace it with clear glass or an IR-pass filter); (2) **filter the lens** — an **IR-pass (visible-blocking) filter** (e.g. R72 / 720 nm) on a stock camera, at the cost of long exposures since most IR is being thrown away by the still-present internal cut filter.

• **what changes — the look**: **foliage glows bright white** (chlorophyll/leaf cell structure reflects NIR strongly) — the classic **"Wood effect"** (after R. W. Wood, 1910); **skin** goes smooth and waxy (NIR penetrates the surface, veins show); **blue sky darkens** and **haze is cut** (less Rayleigh scattering at longer wavelengths → see further); **water** goes inky black.

• **false-color IR**: since NIR is invisible, render it by **mapping bands to visible channels** (e.g. NIR→red, red→green, …) — the surreal magenta-foliage "Aerochrome" look; a channel-remapping choice, not a measurement.

• **uses beyond art**: **forensics & art conservation** (read faded ink, underdrawings, alterations beneath paint); **agriculture / remote sensing** — **NDVI** $=(\text{NIR}-\text{Red})/(\text{NIR}+\text{Red})$, the vegetation-health index that exploits the foliage-glows-in-NIR effect; **astronomy** (NIR pierces dust, redshifted light — e.g. JWST); and as an **extra channel** for computational photography (NIR-assisted denoising / dehazing / dark flash, cross-ref).