Recent Advances in Transparent Electrodes and Their Multimodal Sensing Applications.

This review examines the recent advancements in transparent electrodes and their crucial role in multimodal sensing technologies. Transparent electrodes, notable for their optical transparency and electrical conductivity, are revolutionizing sensors by enabling the simultaneous detection of diverse physical, chemical, and biological signals. Materials like graphene, carbon nanotubes, and conductive polymers, which offer a balance between optical transparency, electrical conductivity, and mechanical flexibility, are at the forefront of this development. These electrodes are integral in various applications, from healthcare to solar cell technologies, enhancing sensor performance in complex environments. The paper addresses challenges in applying these electrodes, such as the need for mechanical flexibility, high optoelectronic performance, and biocompatibility. It explores new materials and innovative techniques to overcome these hurdles, aiming to broaden the capabilities of multimodal sensing devices. The review provides a comparative analysis of different transparent electrode materials, discussing their applications and the ongoing development of novel electrode systems for multimodal sensing. This exploration offers insights into future advancements in transparent electrodes, highlighting their transformative potential in bioelectronics and multimodal sensing technologies.

Gaze location prediction for broadcast football video.

The sensitivity of the human visual system decreases dramatically with increasing distance from the fixation location in a video frame. Accurate prediction of a viewer's gaze location has the potential to improve bit allocation, rate control, error resilience, and quality evaluation in video compression. Commercially, delivery of football video content is of great interest because of the very high number of consumers. In this paper, we propose a gaze location prediction system for high definition broadcast football video. The proposed system uses knowledge about the context, extracted through analysis of a gaze tracking study that we performed, to build a suitable prior map. We further classify the complex context into different categories through shot classification thus allowing our model to prelearn the task pertinence of each object category and build the prior map automatically. We thus avoid the limitation of assigning the viewers a specific task, allowing our gaze prediction system to work under free-viewing conditions. Bayesian integration of bottom-up features and top-down priors is finally applied to predict the gaze locations. Results show that the prediction performance of the proposed model is better than that of other top-down models that we adapted to this context.