International audienc

This article presents the design and implementation of a handheld Augmented Reality (AR) system called Mobile Augmented Reality Touring System (M.A.R.T.S). The results of experiments conducted during museum visits using this system are also described. These experiments aim at studying how such a tool can transform the visitor's learning experience by comparing it to two widely used museum systems. First, we present the museum's learning experience and a related model which emerged from the state of the art. This model consists of two types of activity experienced by the observer of a work of art: sensitive and analytical. Then, we detail M.A.R.T.S architecture and implementation. Our empirical study highlights the fact that AR can direct visitors' attention by emphasizing and superimposing. Its magnifying and sensitive effects are well perceived and appreciated by visitors. The obtained results reveal that M.A.R.T.S contributes to a worthwhile learning experience.

No abstract

Automatic detection of animals that have strayed into human inhabited areas has important security and road safety applications. This paper attempts to solve this problem using deep learning techniques from a variety of computer vision fields including object detection, tracking, segmentation and edge detection. Several interesting insights are elicited into transfer learning while adapting models trained on benchmark datasets for real world deployment. Empirical evidence is presented to demonstrate the inability of detectors to generalize from training images of animals in their natural habitats to deployment scenarios of man-made environments. A solution is also proposed using semi-automated synthetic data generation for domain specific training. Code and data used in the experiments are made available to facilitate further work in this domain.

Very few benchmark exists for assessing pattern detection and recognition in streams in general and for gesture processing in particular. We propose a dedicated benchmark based on the construction of isolated gestures and gesture sequences datasets. This benchmark is associated to a general assessment methodology for streaming processing which first consists in labelling the stream according to some heuristics (that can be optimized on training data) and then aligning the ground truth labelling with the predicted one. 6 pattern recognition models (including DTW, KDTW, HMM, HCRF and SVM) have been accordingly evaluated using this benchmark. It turns out that the regularized kernelized version of DTW measure (KDTW) associated to a SVM is quite efficient, comparatively to the other models, for detecting and recognizing continuous gestures in streams.

No abstract

Very few benchmark exists for assessing pattern detection and recognition in streams in general and for gesture processing in particular. We propose a dedicated benchmark based on the construction of isolated gestures and gesture sequences datasets. This benchmark is associated to a general assessment methodology for streaming processing which first consists in labelling the stream according to some heuristics (that can be optimized on training data) and then aligning the ground truth labelling with the predicted one. 6 pattern recognition models (including DTW, KDTW, HMM, HCRF and SVM) have been accordingly evaluated using this benchmark. It turns out that the regularized kernelized version of DTW measure (KDTW) associated to a SVM is quite efficient, comparatively to the other models, for detecting and recognizing continuous gestures in streams.

Abstract. Although man has become sedentary over time, his wish to travel the world remains as strong as ever. The aim of this paper is to show how techniques based on imagery and Augmented Reality (AR) can prove to be of great help when discovering a new urban environment and observing the evolution of the natural environment. The study's support is naturally the Smartphone which in just a few years has become our most familiar device, which we take with us practically everywhere we go in our daily lives. In this chapter, we discuss technical issues of augmented reality. We deal especially with building recognition. Our building recognition method is based on an efficient hybrid approach, which combines the potentials of SURF features points and features lines. Our method relies on ANNS (Approximate Nearest Neighbors Search) approach, described by Muja et al. [11]. ANNS approaches are known for their speed but they are less accurate than linear algorithms. To assure an optimal trade-off between speed and accuracy, the proposed method performs a filtering step on the top of the Approximate Nearest Neighbors Search. At the last step, our method calls Hausdorff measure [15] with line models.