Ex-Apple Engineers Reveal AI Photography
Ex-Apple engineers discussed how AI is moving from enhancement to creative partner in smartphone photography, enabling real-time scene optimization and computational bokeh. They noted AI isn't just improving photos but "helping people express ideas that weren't possible with traditional cameras."
The engineers behind the iPhone's iconic Portrait Mode, Ziv Attar and Tom Bishop, are now leading the charge in AI photography with their startup, GLASS Imaging. Their goal is to use AI to overcome the physical limitations of small smartphone camera sensors and lenses. At its core, computational photography uses software to create a better image than the camera's hardware could capture on its own. This involves techniques like merging multiple exposures to improve dynamic range and reduce noise, which is especially noticeable in low-light photos. GLASS Imaging is developing a neural network that processes raw data directly from the camera sensor. This end-to-end AI solution replaces numerous traditional processing steps like de-mosaicing, sharpening, and color correction with a single, more powerful process. This AI-driven approach allows for more sophisticated scene understanding. The software can recognize specific elements like faces, skies, or food and apply targeted adjustments, ensuring, for example, that skin tones are accurate while the background is vibrant.
