Panasonic Holdings Co. Ltd (Panasonic HD) has announced the development of a new AI algorithm system introducing a new classification algorithm that can handle multimodal distribution. As of writing this article, there are no real-world photographic implementations to speak of, but the tech itself is intriguing. Is a new Panasonic AI autofocus system coming? AI, deep learning, and subject recognition Subject recognition algorithms aren’t new. Their basic forms have been with us since before the dawn of mirrorless. Readers may remember Canon’s early attempts with DSLR-based iTR systems and how early camcorders and compact cameras could adjust settings based on shooting conditions, and there may be other instances where photographic devices used different sorts of recognition. The evolution continued and took an exponential leap forward with the mirrorless revolution. Autofocus subject recognition, LUMIX G9 II. Image credit: Panasonic UK Mirrorless and AI – a perfect match Mirrorless technology isn’t new. Digital cameras with some forms of live view have been around for over 20 years now. The significant development in recent years came with the exponential increase in data transfer and sensor readout speed. Cameras these days crunch through incredible amounts of data. AI and its more common variant – deep learning, thrive on big data. This convergence of a camera’s ability to provide a constant flow of live data and pre-learned algorithmic data enables significantly better interpretation of the scene. Subject recognition was just a matter of time, and we can now find it in most, if not all,...
Published By: CineD - Thursday, 11 January, 2024