Core Image Labo vs YAME

Core Image Labo is a developer tool for testing a Core Image filter chain. CIFilters with scalar input parameters are currently supported. CIFilters that require a background or target image are also supported. All other CIFilters are optionally listed for reference. ・Filters can be enabled and disabled by tapping the filter name. ・Filters can be added, reordered, and removed. ・Multiple filters of the same type can be added. ・Export the final image and JSON filter chain for later use. Core Image Labo is open source, so feel free to contribute bug fixes, new features, a better filter support.

YAME is an intelligent assistant app built upon the latest on-device Vision Large Language Model. Designed for the visually impaired, YAME helps users perceive and understand the world through state-of-the-art vision-language AI. Key Features • Image Description: Understand your surroundings with detailed visual descriptions. • Facial Expression Recognition: Detect emotions from facial expressions to aid social interaction. • Traffic Signal Detection: Identify traffic light states for safer navigation. • Text-to-Speech (TTS): All results are read aloud instantly, providing seamless audio feedback. Private by Design – Offline AI All processing is done 100% on-device. No internet connection is required, and no images or personal data are ever uploaded or shared. YAME is built with privacy at its core. Device Requirements Due to the high memory demand of on-device AI inference, YAME is recommended for use on the following devices: Compatible iPhones: • iPhone 12 Pro / 12 Pro Max • iPhone 13 Pro / 13 Pro Max • iPhone 14 / 14 Plus / 14 Pro / 14 Pro Max • iPhone 15 / 15 Plus / 15 Pro / 15 Pro Max • iPhone 16 / 16 Plus / 16 Pro / 16 Pro Max Compatible iPads: • iPad Air (5th Gen, M1 chip) • iPad Pro 11-inch (3rd Gen and later, M1 or M2 chip) • iPad Pro 12.9-inch (5th Gen and later, M1 or M2 chip) Note: Devices with less than 6GB RAM (e.g., iPhone 12 or earlier, non-Pro iPads) may not run the model reliably. For best performance and stability, please use a supported device.