Grenoble Institute of Technology

About: Grenoble Institute of Technology is a education organization based out in Grenoble, France. It is known for research contribution in the topics: Hyperspectral imaging & Geology. The organization has 3427 authors who have published 5345 publications receiving 137158 citations. The organization is also known as: Grenoble INP.

Enhanced Patterns of Oriented Edge Magnitudes for Face Recognition and Image Matching

Abstract: A good feature descriptor is desired to be discriminative, robust, and computationally inexpensive in both terms of time and storage requirement. In the domain of face recognition, these properties allow the system to quickly deliver high recognition results to the end user. Motivated by the recent feature descriptor called Patterns of Oriented Edge Magnitudes (POEM), which balances the three concerns, this paper aims at enhancing its performance with respect to all these criteria. To this end, we first optimize the parameters of POEM and then apply the whitened principal-component-analysis dimensionality reduction technique to get a more compact, robust, and discriminative descriptor. For face recognition, the efficiency of our algorithm is proved by strong results obtained on both constrained (Face Recognition Technology, FERET) and unconstrained (Labeled Faces in the Wild, LFW) data sets in addition with the low complexity. Impressively, our algorithm is about 30 times faster than those based on Gabor filters. Furthermore, by proposing an additional technique that makes our descriptor robust to rotation, we validate its efficiency for the task of image matching.

High-Capacity Chipless RFID Tag Insensitive to the Polarization

Abstract: Designing a reader for chipless RFID is a hard task since both the polarization and operating frequency agility have to be implemented. The new tag design proposed in this paper is polarization independent, making the design of the reader easier since only linear polarization is needed to detect the tag. The proposed chipless tag is based on multiple circular ring patch resonators. The coding capacity of this tag reaches 19 bits within a compact surface of cm . Further, the frequency band is within 3.1 to 10.6 GHz to be compliant with FCC and ECC regulations for UWB. This new design is experimentally validated in the frequency domain using bi-static measurement set-up. Both amplitude and group delay responses of the tag are investigated and carried out.

Quantum effect on thermally activated glide of dislocations.

Abstract: Crystal plasticity involves the motion of dislocations under stress. So far, atomistic simulations of this process have predicted Peierls stresses, the stress needed to overcome the crystal resistance in the absence of thermal fluctuations, of more than twice the experimental values, a discrepancy best-known in body-centred cubic crystals. Here we show that a large contribution arises from the crystal zero-point vibrations, which ease dislocation motion below typically half the Debye temperature. Using Wigner's quantum transition state theory in atomistic models of crystals, we found a large decrease of the kink-pair formation enthalpy due to the quantization of the crystal vibrational modes. Consequently, the flow stress predicted by Orowan's law is strongly reduced when compared with its classical approximation and in much closer agreement with experiments. This work advocates that quantum mechanics should be accounted for in simulations of materials and not only at very low temperatures or in light-atom systems.

Hyperspectral Remote Sensing Image Classification Based on Rotation Forest

Abstract: In this letter, an ensemble learning approach, Rotation Forest, has been applied to hyperspectral remote sensing image classification for the first time. The framework of Rotation Forest is to project the original data into a new feature space using transformation methods for each base classifier (decision tree), then the base classifier can train in different new spaces for the purpose of encouraging both individual accuracy and diversity within the ensemble simultaneously. Principal component analysis (PCA), maximum noise fraction, independent component analysis, and local Fisher discriminant analysis are introduced as feature transformation algorithms in the original Rotation Forest. The performance of Rotation Forest was evaluated based on several criteria: different data sets, sensitivity to the number of training samples, ensemble size and the number of features in a subset. Experimental results revealed that Rotation Forest, especially with PCA transformation, could produce more accurate results than bagging, AdaBoost, and Random Forest. They indicate that Rotation Forests are promising approaches for generating classifier ensemble of hyperspectral remote sensing.

Chirality-Induced asymmetric magnetic nucleation in Pt/Co/AlOx ultrathin microstructures.

Abstract: The nucleation of reversed magnetic domains in Pt/Co/AlO(x) microstructures with perpendicular anisotropy was studied experimentally in the presence of an in-plane magnetic field. For large enough in-plane field, nucleation was observed preferentially at an edge of the sample normal to this field. The position at which nucleation takes place was observed to depend in a chiral way on the initial magnetization and applied field directions. A quantitative explanation of these results is proposed, based on the existence of a sizable Dzyaloshinskii-Moriya interaction in this sample. Another consequence of this interaction is that the energy of domain walls can become negative for in-plane fields smaller than the effective anisotropy field.