Search results

Talks:Getting the Most out of Your Image Signal Processor, Dr. Claudio Greco, Image Science Director, DXOMARK

The continuous scaling of pixel size is increasingly constrained by the minimum area needed for pixel circuitry, while emerging applications demand simultaneously higher resolution, wider dynamic range, lower noise, and enhanced optical functionality. To address these challenges, imec is exploring a new image sensor architecture that combines back-end-of-line (BEOL) integrated Indium-Gallium-Zinc-Oxide (IGZO) thin-film transistors with advanced meta-optical components.

Ambisonics offers a signal-independent and mathematically rigorous framework for representing spatial audio. Widely adopted in industry and deeply studied in academia, Ambisonics models the sound field around a listener’s head using spherical harmonics. This makes it particularly well suited for headphone-based applications such as music playback, virtual reality (VR), and augmented reality (AR). This webinar provides an overview of Ambisonics recording and encoding.

Modern imaging systems are evolving beyond image capture and display toward integrated perception systems that help people understand and act in the physical world. Advances in sensors, computational imaging, computer vision, foundation models, and real-time graphics are enabling new forms of augmented vision that can enhance situational awareness across applications ranging from transportation and industrial operations to defense and aviation.

Deep learning has become the de facto standard method in most image processing problems. In the past few years, deep learning algorithms have met and exceeded human-level performance in image recognition. Nevertheless, training deep learning networks on a large data set remains very challenging. The sheer amount of computation needed to train a convolutional neural network can take months on large data sets. Combining that with the black art of hyperparameter tuning, the community desperately needs tools to help train deep learning networks on multiple servers with multiple GPUs.

There is a fast development of different machine learning methods – for object classification, tracking, action recognition and other tasks with multiple types of data – from images and videos to data from wireless sensor networks. Autonomous image and video analytics face a number of challenges due to the huge volumes of data that sensors provide, the changeable environmental conditions and other factors. However, it is important to know when the methods work well and when they are not reliable, e.g. how much could we trust the obtained results?

This talk is about conversational AI and how far are we from being able to talk to AI in natural languages. The talk will touch upon modelling and implementation of the AI embodiment in a photorealistic manner. It will explain machine learning, computer vision and Generative AI methods developed to this end and used for photorealistic AI Buddies generation. The talk will end in discussion of possible applications and misuses of this technology, including Deep Fakes.

The importance of large databases of images in modern computer vision is hard to overstate. In the last decade, the scale of vision datasets has been increasing at a rapid pace, but these datasets come with costs: image collection and data annotation are expensive. To counter these costs, interest has surged in unsupervised learning, as it avoids the curation efforts. But images are still hard to collect. An alternative to real images is to use simulations with graphics engines, but content creation is also costly.