Profil du membre

One of the best ways to build better vision models is to train large models on big datasets. However, the process of building such datasets is often costly and tedious. With the ever increasing adoption of Mixed Reality in professional settings, and with the performance improvements of headsets in recent years, we see an opportunity for a tool that combines data collection and annotation in a single process, and that leverages both RGB and depth data provided by the headset sensors. Moreover, assisting machine learning predictive models with user inputs through natural mixed reality interactions is a promising prospective for human-artificial intelligence interactions. In this paper, we present MRLabeling, an application developped for the Microsoft Hololens 2 that allows the easy creation and annotation of datasets directly in Mixed Reality. We first describe the design of the system, the way 3D bounding boxes drawn by the user are projected in 2D to create annotated images on the fly, and the use of segmentation algorithms to go beyond bounding boxes. After that, we explore the use of depth data, and the current limitations of the system, as well as avenues for future work.

Mixed reality has made its first step towards democratization in 2017 with the launch of a first generation of commercial devices. As a new medium, one of the challenges is to develop interactions using its endowed spatial awareness and body tracking. More specifically, at the crossroad between artificial intelligence and human-computer interaction, the goal is to go beyond the Window, Icon, Menu, Pointer (WIMP) paradigm humans are mainly using on desktop computer. Hand interactions either as a standalone modality or as a component of a multimodal modality are one of the most popular and supported techniques across mixed reality prototypes and commercial devices. In this context, this paper presents scoping literature review of hand interactions in mixed reality. The goal of this review is to identify the recent findings on hand interactions about their design and the place of artificial intelligence in their development and behavior. This review resulted in the highlight of the main interaction techniques and their technical requirements between 2017 and 2022 as well as the design of the Metaphor-behavior taxonomy to classify those interactions.