No abstract

Speaker recognition is an important classification task, which can be solved using several approaches. Although building a speaker recognition model on a closed set of speakers under neutral speaking conditions is a well-researched task and there are solutions that provide excellent performance, the classification accuracy of developed models significantly decreases when applying them to emotional speech or in the presence of interference. Furthermore, deep models may require a large number of parameters, so constrained solutions are desirable in order to implement them on edge devices in the Internet of Things systems for real-time detection. The aim of this paper is to propose a simple and constrained convolutional neural network for speaker recognition tasks and to examine its robustness for recognition in emotional speech conditions. We examine three quantization methods for developing a constrained network: floating-point eight format, ternary scalar quantization, and binary scalar quantization. The results are demonstrated on the recently recorded SEAC dataset.

U sintezi ekspresivnog govora važno je generisati emocionalno obojen govor koji odražava kompleksnost emocionalnih stanja. Brojni TTS sistemi emocije u sintetizovanom govoru modeluju u vidu diskretnih skupova, ali tek kada se uzmu u obzir i varijacije koje postoje unutar emotivnih stanja, generisani govor može biti nalik ljudskom. Ovaj rad obuhvata teorijsku analizu i poređenje dva inovativna sistema za sintezu ekspresivnog govora koji kompleksnost emocija modeluju u vidu kontinualnih vektora kojima je moguće manipulisati. Rezultati pokazuju da je pristup zasnovan na t-SNE embedding vektorima primjenljiv samo u slučaju specifičnih baza podataka, dok je drugi pristup, zasnovan na interpolaciji tačaka u embedding prostoru multi-speaker, multi-style modela, opštiji, ali zahtijeva dodatnu analizu.

Background and Objectives: Foot deformities are the basis of numerous disorders of the locomotor system. An optimized method of classification of foot deformities would enable an objective identification of the type of deformity since the current assessment methods do not show an optimal level of objectivity and reliability. The acquired results would enable an individual approach to the treatment of patients with foot deformities. Thus, the goal of this research study was the development of a new, objective model for recognizing and classifying foot deformities with the application of machine learning, by labeling baropodometric analysis data using computer vision methods. Materials and Methods: In this work, data from 91 students of the Faculty of Medicine and the Faculty of Sports and Physical Education, University of Novi Sad were used. Measurements were determined by using a baropodometric platform, and the labelling process was carried out in the Python programming language, using functions from the OpenCV library. Segmentation techniques, geometric transformations, contour detection and morphological image processing were performed on the images, in order to calculate the arch index, a parameter that gives information about the type of the foot deformity. Discussion: The foot over which the entire labeling method was applied had an arch index value of 0.27, which indicates the accuracy of the method and is in accordance with the literature. On the other hand, the method presented in our study needs further improvement and optimization, since the results of the segmentation techniques can vary when the images are not consistent. Conclusions: The labeling method presented in this work provides the basis for further optimization and development of a foot deformity classification system.

No abstract