Combining Haptic and Bang-Bang Braking Actions for Passive Robotic Walker Path Following.

Robotic walkers are a promising solution for physical and cognitive support to older adults. This paper proposes a low cost path following strategy combining the advantages of a simple mechanical braking guidance, such as safety, passivity, and a low cost, and the ones of a vibrotactile haptic guidance, such as comfort and portability. The user is guided by providing indications on the directions of motion using the haptic interface so that he/she can autonomously and comfortably follow the planned path. However, whenever the user significantly departs from the path (for instance, s/he gets too close to obstacles), the braking system kicks in to safely steer the user back along the proper direction. The formal correctnesses of the hybrid strategy ruling the combination of the two guidance systems is proved theoretically. Moreover, a comprehensive experimental study with users aged 64 to 100, including also psychological evaluations, has been performed. The hybrid combination of the braking and the haptic guidance systems is shown to outperform the two individual approaches in isolation. The combination of the two retains the same level of the users' perceived comfort typical of the haptic-only guidance while ensuring the adequate path following performance typical of the braking-only guidance. In particular, the combined approach produces a mean path following error equal to 41% of the mean path following error ensured by the haptic-only approach. Conversely, thanks to the haptic feedback, the combined approach halves the activation time and the number of interventions needed in the braking-only approach.

Unsupervised learning of categorical segments in image collections.

Which one comes first: segmentation or recognition? We propose a unified framework for carrying out the two simultaneously and without supervision. The framework combines a flexible probabilistic model, for representing the shape and appearance of each segment, with the popular "bag of visual words" model for recognition. If applied to a collection of images, our framework can simultaneously discover the segments of each image and the correspondence between such segments, without supervision. Such recurring segments may be thought of as the "parts" of corresponding objects that appear multiple times in the image collection. Thus, the model may be used for learning new categories, detecting/classifying objects, and segmenting images, without using expensive human annotation.

Metrological validation for 3D modeling of dental plaster casts.

The contribution of this paper is twofold: (1) it presents an automatic 3D modeling technique and (2) it advances a procedure for its metrological evaluation in the context of a medical application, the 3D modeling of dental plaster casts. The motivation for this work is the creation of a "virtual gypsotheque" where cumbersome dental plaster casts can be replaced by numerical 3D models, thereby alleviating storage and access problems and allowing dentists and orthodontists the use of novel and unprecedented software tools for their medical evaluations. Modeling free-form surfaces of anatomical interest is an intriguing mixture of open issues concerning 3D modeling, geometrical metrology, and medicine. Of general interest is both the fact that a widespread use of 3D modeling in non-engineering applications requires automatic procedures of the kind presented in this work and the adopted validation paradigm for free-form surfaces, rather useful for practical purposes. In this latter respect, the metrological analysis we advance is the first seminal attempt in the field of 3D modeling and can be readily extended to contexts other than the medical one discussed in this paper.

Automatic 3-d modeling of textured cultural heritage objects.

A widespread use of three-dimensional (3-D) models in cultural heritage application requires low cost equipment and technically simple modeling procedures. In this context, methods for automatic 3-D modeling of textured objects will play a central role. Such methods need fully automatic techniques for 3-D views registration and for the removal of texture artifacts. This paper proposes a contribution in this direction based on image processing approaches. The proposed procedure is very robust and simple. It does not require special equipment or skill in order to make textured 3-D models. The results of this paper, originally conceived to address the costs issues of cultural heritage modeling, can be profitably exploited also in other modeling applications.