Understanding the timing of economic feasibility: The case of input interfaces for human-computer interaction

Abstract

The objective of this paper is to understand when new types of input interfaces for Human–Computer Interaction (HCI) such as Natural User Interfaces (NUI) (e.g., speech and gesture) and Direct Neural Interfaces (DNI), or combinations of them, might become technologically and economically feasible. This problem is addressed by analyzing the performance trajectories of key components in these HCI systems. In the case of speech interfaces, we observe that microphones and automated speech recognition systems are no longer experiencing rapid improvements along key dimensions of performance, which inhibits their technical and economic feasibility. On the other hand, 2D image sensors and depth sensors, which constitute the core components of gesture interfaces, are continuing to improve at a significant rate in terms of characteristics like spatial resolution, pixel sensitivity, and depth resolution. When coupled with the exponential improvements in the memory and processing power of computing systems, the above improvements in image sensors are enabling gesture-based natural user interfaces to reach acceptable levels of technical performance and economic feasibility. Similarly, simultaneous improvement in the spatial and temporal resolution of non-invasive brain scanning technologies is likely to accelerate the development of direct neural interfaces (DNI). However, a number of challenging obstacles such as lack of robust magnetic shielding systems, high cost, and poor usability continue to hinder the economic feasibility of DNI systems.