Hand gesture recognition for human computer interaction : a comparative study of different image features

Hand gesture recognition for human computer interaction, being a natural way of human computer interaction, is an area of active research in computer vision and machine learning. This is an area with many different possible applications, giving users a simpler and more natural way to communicate with robots/systems interfaces, without the need for extra devices. So, the primary goal of gesture recognition research is to create systems, which can identify specific human gestures and use them to convey information or for device control. For that, vision-based hand gesture interfaces require fast and extremely robust hand detection, and gesture recognition in real time. In this study we try to identify hand features that, isolated, respond better in various situations in human-computer interaction. The extracted features are used to train a set of classifiers with the help of RapidMiner in order to find the best learner. A dataset with our own gesture vocabulary consisted of 10 gestures, recorded from 20 users was created for later processing. Experimental results show that the radial signature and the centroid distance are the features that when used separately obtain better results, with an accuracy of 91% and 90,1% respectively obtained with a Neural Network classifier. These to methods have also the advantage of being simple in terms of computational complexity, which make them good candidates for real-time hand gesture recognition.

Hand gesture recognition system based in computer vision and machine learning

"Lecture notes in computational vision and biomechanics series, ISSN 2212-9391, vol. 19"

Generic system for human-computer gesture interaction: applications on sign language recognition and robotic soccer refereeing

Hand gestures are a powerful way for human communication, with lots of potential applications in the area of human computer interaction. Vision-based hand gesture recognition techniques have many proven advantages compared with traditional devices, giving users a simpler and more natural way to communicate with electronic devices. This work proposes a generic system architecture based in computer vision and machine learning, able to be used with any interface for human-computer interaction. The proposed solution is mainly composed of three modules: a pre-processing and hand segmentation module, a static gesture interface module and a dynamic gesture interface module. The experiments showed that the core of visionbased interaction systems could be the same for all applications and thus facilitate the implementation. For hand posture recognition, a SVM (Support Vector Machine) model was trained and used, able to achieve a final accuracy of 99.4%. For dynamic gestures, an HMM (Hidden Markov Model) model was trained for each gesture that the system could recognize with a final average accuracy of 93.7%. The proposed solution as the advantage of being generic enough with the trained models able to work in real-time, allowing its application in a wide range of human-machine applications. To validate the proposed framework two applications were implemented. The first one is a real-time system able to interpret the Portuguese Sign Language. The second one is an online system able to help a robotic soccer game referee judge a game in real time.

Generic system for human-computer gesture interaction

Hand gestures are a powerful way for human communication, with lots of potential applications in the area of human computer interaction. Vision-based hand gesture recognition techniques have many proven advantages compared with traditional devices, giving users a simpler and more natural way to communicate with electronic devices. This work proposes a generic system architecture based in computer vision and machine learning, able to be used with any interface for humancomputer interaction. The proposed solution is mainly composed of three modules: a pre-processing and hand segmentation module, a static gesture interface module and a dynamic gesture interface module. The experiments showed that the core of vision-based interaction systems can be the same for all applications and thus facilitate the implementation. In order to test the proposed solutions, three prototypes were implemented. For hand posture recognition, a SVM model was trained and used, able to achieve a final accuracy of 99.4%. For dynamic gestures, an HMM model was trained for each gesture that the system could recognize with a final average accuracy of 93.7%. The proposed solution as the advantage of being generic enough with the trained models able to work in real-time, allowing its application in a wide range of human-machine applications.

Para cá e para lá dos muros: relações familiares na interface entre o interior e o exterior da prisão

Tese de Doutoramento em Sociologia

The role of social cognition and motor imagery in gesture processing: a patients' study

Dissertação de mestrado integrado em Psicologia

Effect of the acoustic impedance in ultrasonic emitter transducers using digital modulations

In an underwater environment it is difficult to implement solutions for wireless communications. The existing technologies using electromagnetic waves or lasers are not very efficient due to the large attenuation in the aquatic environment. Ultrasound reveals a lower attenuation, and thus has been used in underwater long-distance communications. The much slower speed of acoustic propagation in water (about 1500 m/s) compared with that of electromagnetic and optical waves, is another limiting factor for efficient communication and networking. For high data-rates and real-time applications it is necessary to use frequencies in the MHz range, allowing communication distances of hundreds of meters with a delay of milliseconds. To achieve this goal, it is necessary to develop ultrasound transducers able to work at high frequencies and wideband, with suitable responses to digital modulations. This work shows how the acoustic impedance influences the performance of an ultrasonic emitter transducer when digital modulations are used and operating at frequencies between 100 kHz and 1 MHz. The study includes a Finite Element Method (FEM) and a MATLAB/Simulink simulation with an experimental validation to evaluate two types of piezoelectric materials: one based on ceramics (high acoustic impedance) with a resonance design and the other based in polymer (low acoustic impedance) designed to optimize the performance when digital modulations are used. The transducers performance for Binary Amplitude Shift Keying (BASK), On-Off Keying (OOK), Binary Phase Shift Keying (BPSK) and Binary Frequency Shift Keying (BFSK) modulations with a 1 MHz carrier at 125 kbps baud rate are compared.

Sistema de visão robótico baseado no sensor kinect para orientação no espaço e contorno de obstáculos

Dissertação de mestrado em Engenharia Eletrónica Industrial e Computadores (área de especialização em Robótica)

Technical requirements for the use of C&DW recycled aggregates in municipal and rural roads and trenches

The use of construction and demolition waste (C&DW) in the construction industry is an important contribution to attain sustainability in the sector. The roads are among the civil engineering works which can use larger quantities of C&DW recycled aggregates. In Portugal, the limit values for the properties of C&DW recycled aggregates that can be used in the roads of Portuguese Road Network are defined by two Laboratório Nacional de Engenharia Civil (LNEC) technical specifications (TS), in accordance to Portuguese Decree-law no. 46/2008 of May 12th. Municipal and rural roads and trenches have specific characteristics that can enable the use of C&DW of lower quality than those required by existing LNEC TS, and even then ensuring an adequate performance. However, given the absence of specific regulation for those applications, the Portuguese Environment Agency requires compliance with the existing LNEC TS, which represents an obstacle to recycling a significant part of the C&DW, in particular at a local government level. This paper presents guidelines for the recycling of C&DW in municipal and rural roads and in trenches, which could be considered in a new forthcoming LNEC TS. In the preparation of the guidelines, the bibliography collected and analysed, the information gathered from the application of C&DW in a municipal and rural roads of a Portuguese municipality and in the roadways of a Portuguese resort, and the results of laboratory tests carried out on samples collected in the Portuguese municipality were taken into consideration.

Hand gesture recognition system based in computer vision and machine learning: Applications on human-machine interaction

Tese de Doutoramento em Engenharia de Eletrónica e de Computadores

Ancestry of the Brazilian TP53 c.1010G>A (p.Arg337His, R337H) founder mutation : clues from haplotyping of short tandem repeats on Chromosome 17p

Rare germline mutations in TP53 (17p13.1) cause a highly penetrant predisposition to a specific spectrum of early cancers, defining the Li-Fraumeni Syndrome (LFS). A germline mutation at codon 337 (p.Arg337His, c1010G>A) is found in about 0.3% of the population of Southern Brazil. This mutation is associated with partially penetrant LFS traits and is found in the germline of patients with early cancers of the LFS spectrum unselected for familial his- tory. To characterize the extended haplotypes carrying the mutation, we have genotyped 9 short tandem repeats on chromosome 17p in 12 trios of Brazilian p.Arg337His carriers. Results confirm that all share a common ancestor haplotype of Caucasian/Portuguese-Ibe- ric origin, distant in about 72–84 generations (2000 years assuming a 25 years intergenera- tional distance) and thus pre-dating European migration to Brazil. So far, the founder p. Arg337His haplotype has not been detected outside Brazil, with the exception of two resi- dents of Portugal, one of them of Brazilian origin. On the other hand, increased meiotic recombination in p.Arg337His carriers may account for higher than expected haplotype diversity. Further studies comparing haplotypes in populations of Brazil and of other areas of Portuguese migration are needed to understand the historical context of this mutation in Brazil.

Impact of TGF-ß1 -509C/T and 869T/C polymorphisms on glioma risk and patient prognosis

Transforming growth factor beta (TGF-ß) plays an important role in carcinogenesis. Two polymorphisms in the TGF-ß1 gene (-509C/T and 869T/C) were described to influence susceptibility to gastric and breast cancers. The 869T/C polymorphism was also associated with overall survival in breast cancer patients. In the present study, we investigated the relevance of these TGF-ß1 polymorphism in glioma risk and prognosis. A case-control study that included 114 glioma patients and 138 cancer-free controls was performed. Single nucleotide polymorphisms (SNPs) were evaluated by polymerase chain reaction followed by restriction fragment length polymorphism (PCR-RFLP). Univariate and multivariate logistic regression analyses were used to calculate odds ratio (OR) and 95 % confidence intervals (95 % CI). The influence of TGF-ß1 -509C/T and 869T/C polymorphisms on glioma patient survival was evaluated by a Cox regression model adjusted for patients' age and sex and represented in Kaplan-Meier curves. Our results demonstrated that TGF-ß1 gene polymorphisms -509C/T and 869T/C are not significantly associated with glioma risk. Survival analyses showed that the homozygous -509TT genotype associates with longer overall survival of glioblastoma (GBM) patients when compared with patients carrying CC + CT genotypes (OR, 2.41; 95 % CI, 1.06-5.50; p = 0.036). In addition, the homozygous 869CC genotype is associated with increased overall survival of GBM patients when compared with 869TT + TC genotypes (OR, 2.62; 95 % CI, 1.11-6.17; p = 0.027). In conclusion, this study suggests that TGF-ß1 -509C/T and 869T/C polymorphisms are not significantly associated with risk for developing gliomas but may be relevant prognostic biomarkers in GBM patients.

Rheological characterization of EVA and HDPE polymer modified bitumens under large deformation at 20 º C

Large amplitude oscillatory shear (LAOS) coupled with Fourier transform rheology (FTR) was used for the first time to characterize the large deformation behavior of selected bituminous binders at 20 C. Two polymer modified bitumens (PMB) containing recycled EVA and HDPE and two unmodified bitumens were tested with LAOS-FTR. The LAOS-FTR response of all binders was compared at same frequency, at same Deborah number (by tuning the frequency to the relaxation time of each binder) and at same phase shift angle d (by tuning the frequency to the one corresponding to d = 50 in the SAOS response of each sample). In all the approaches, LAOS-FTR results allowed to differentiate between all the nonlinear mechanical characteristics of the tested binders. All binders show LAOS-FTR patterns reminiscent from colloidal dispersions and emulsions. EVA PMB was less prone to strain-induced microstructural changes when compared to HDPE PMB which showed larger values of nonlinear FTR parameters for the range of shear strains tested in LAOS.

High performance screen printable lithium-ion battery cathode ink based on C-LiFePO4

Lithium-ion battery cathodes have been fabricated by screen-printing through the development of CLiFePO4 inks. It is shown that shear thinning polymer solutions in N-methyl-2-pyrrolidone (NMP) with Newtonian viscosity above 0.4 Pa s are the best binders for formulating a cathode paste with satisfactory film forming properties. The paste shows an elasticity of the order of 500 Pa and, after shear yielding, shows an apparent viscosity of the order of 3 Pa s for shear rates corresponding to those used during screen-printing. The screen-printed cathode produced with a thickness of 26 mm shows a homogeneous distribution of the active material, conductive additive and polymer binder. The total resistance and diffusion coefficient of the cathode are 450 V and 2.5 10 16cm2 s 1, respectively. The developed cathodes show an initial discharge capacity of 48.2 mAh g 1 at 5C and a discharge value of 39.8 mAh g 1 after 50 cycles. The capacity retention of 83% represents 23% of the theoretical value (charge and/or discharge process in twenty minutes), demonstrating the good performance of the battery. Thus, the developed C-LiFePO4 based inks allow to fabricate screen-printed cathodes suitable for printed lithium-ion batteries