Adaptive Projection on Suitable Sub-Surface in Indoor Cluttered Environment

Abstract

The advent of portable smart devices like smart phones, digital cameras, personal digital assistants and tablets resulted in emanation of immense visual data that people wish to view and share among each other. But the small screen of such devices poses a limitation on sharing of visual contents. However, the pervasive technology of handheld or pico projectors overcome this limitation of content sharing by allowing projection of media content on a shared medium. Traditional projection environments make use of projectors in conjunction with custom-built projection screens. Such scenarios typically employ a static setting and thus projectors and projection surfaces are immobile. However, with the availability of pocket-able pico-projectors, the mode of usage of such hand-held projection devices is evolving. Users are now using such devices to project in random environments that may be cluttered and not built for projection. This poses a serious challenge since the entire surface may not be suitable for projection. The users have to search for a suitable surface on which the projection is to be made. This sometimes requires the physical movement of the arm or the person itself to have a confined projection over the surface. Therefore, there is a need to develop a system that will enable pico-projector users to automatically restrict the projection content within a suitable surface in casual indoor settings. The relevant schemes for traditional projectors are not applicable in the context of pico projectors because pico projectors have low resolution and luminosity values. To the best of my knowledge this research is the first work in the field of camera-projector systems aiming to the select the suitable surface for projection and then confining the projection to the selected surface using a binary coded pattern. In particular, there are the following major contributions of this work: Firstly, a novel scheme that allows the projection system to be intelligent enough to confine the projection on a suitable surface that falls within the projector’s field of view. The projection confinement on a suitable surface is performed in two major steps; selecting a suitable surface for projection that lie within the camera-projector view and then confining the projection on a selected surface through homography estimation using the binary coded pattern. For the initial phase, the color image segmentation is performed for the selection of the appropriate surface for projection. The homogenous color segment that is larger in size is selected as a candidate surface for projection. The next step is to confine the projection over the selected surface or the subregion of a surface through homography estimation using binary coded pattern. Secondly, development of a binary coded pattern which enables reliable selection of projection surface by avoiding clutter in the projected area. The developed binary coded pattern outperforms the existing binary coded patterns for the scenario of low resolution cameraprojector pair and hence is particularly suited for pico-projectors that have low resolution and luminosity values. Experiments confirm that the proposed pattern is more immune to detection errors as compared to the state of the art binary coded patterns in the context of mobile grade low resolution pico projectors. A formative study has also been conducted to get an insight regarding the user centric attributes for projection surface selection through a survey based approach. The respondents were asked about the five parameters related to the projection surface that comprises of two attributes each. These attributes are related to the surface size, color, reflectance property, geometry and texture. The simple statistical analysis of the gathered data reveal that users preferred the planar large surfaces for projection that are of the bright color and matte and textureless in nature. Weights are assigned to the surface attributes as suggested by the respondents in the order of (from high to low); size, color, texture, reflectance property and geometry of the surface.