Luminance Gradient Scanner

Every photograph records a scene lit by one or more light sources. This creates a brightness surface — smooth gradients from highlight to shadow that follow the geometry of objects and light direction. The Luminance Gradient Scanner converts the image to grayscale, then applies a 3×3 Sobel operator to compute horizontal and vertical brightness derivatives. The resulting magnitude map shows where brightness changes rapidly (edges, shadows) versus gradually (flat surfaces, skies). When someone composites elements from different photos, the pasted region carries its original lighting, creating abrupt gradient boundaries that don't match the surrounding scene. For noise-level detection, combine with the Noise Analysis Scanner.

Heatmap mode maps gradient magnitude from blue (flat, low gradient) through green and yellow to red (strong edges). Direction mode encodes the gradient angle as hue and magnitude as brightness — consistent lighting appears as a uniform dominant color, while composited regions show a different hue. Grayscale mode shows raw magnitude in black-and-white. Amplification (0.5–3×) scales the gradient for visibility — increase it on low-contrast images, decrease on high-contrast ones. The overlay toggle blends the gradient map at 50% opacity with the original for easier spatial orientation. Use the Re-run button to adjust parameters without re-uploading. For compression-based analysis, try the ELA Scanner.

A single light source produces gradients that flow in a consistent direction across all objects in a scene. The scanner computes the dominant gradient direction using an 8-sector histogram weighted by magnitude. When a person is pasted from a studio photo (lit from the left) into an outdoor scene (lit from above), the Direction mode reveals the mismatch as a contrasting hue in that region. The block-based consistency analysis divides the image into 32×32 blocks and computes the coefficient of variation — a high value with many anomalous blocks strongly suggests compositing. For copy-paste forgery analysis, use Clone Detection. For JPEG quality splice detection, try the JPEG Ghost Scanner.

Forensic investigators use gradient analysis to detect compositing in evidence images — pasted elements with inconsistent lighting are a hallmark of amateurish manipulation. Journalism fact-checkers verify press photos for signs of inserted or removed subjects. Insurance adjusters examine claim photos where objects appear to have been added to a scene with mismatched shadows. Scientific image reviewers check for undisclosed compositing in figure panels. Photographers use it to evaluate lighting setups and identify unwanted light spill. For metadata-level verification, combine with the Authenticity Checker. For a comprehensive forensic workflow, visit the photo forensics hub.