National Repository of Grey Literature 84 records found  1 - 10nextend  jump to record: Search took 0.01 seconds. 
Transluminace plic - hledání solitárních plicních uzlů pomocí IR záření
Votruba, Jiří ; Drahanský, Martin ; Goldmann, Tomáš
Tento článek obsahuje popis vývoje zařízení s metodikou pro transluminaci plicní tkáně s cílem nalezení solitárních plicních uzlů s minimalizací zátěže pro pacienta. Zařízení na prosvětlování plicní tkáně je založeno na infračerveném světle a bronchoskopu (či trokaru), který umožňuje přenos infračerveného záření po průsvitu tkání do kamery. První pokusy byly provedeny na vepřových plicích ex vivo, pokračovali jsme přes vepřové plíce in vivo, nyní se plánuje testování na pacientech.
System for Transmitting Diagnostic Data from the Car
Uhliar, Martin ; Vašíček, Zdeněk (referee) ; Goldmann, Tomáš (advisor)
This diploma thesis deals with the design and implementation at the software and hardware level of the device, which is used for communication and acquisition of diagnostic data from the car control unit. The acquired data can be stored on the device and also sent via LoRa technology and TheThingsNetwork to a database on a remote server. NB-IoT technology is used for data that needs to be obtained in real time. The recorded data is presented in the form of a mobile application, which displays the obtained data to the user and can notify him of a problem with the vehicle. The device communicates with the car via the CAN bus using the OBD2 protocol and is controlled by an ARM microcontroller.
Blood Vessels Detection in the Fingers Area for Liveness Detection
Šulavík, Jan ; Goldmann, Tomáš (referee) ; Rydlo, Štěpán (advisor)
Biometrics as a part of computer technology are recently rising in popularity and various biometric methods are becoming a part of everyday human life. This work in the first place informs the reader on the basics of biometrics in computer technology and further continues to discuss the use of veins in this area as veins are for their uniqueness considered as an ideal human physiological trait for use in biometrics. It is described in what manner is this biometric characteristic being captured. Further more the work describes some image processing methods for extracting biometric data from the input images that are later used. These methods were also implemented and the details of this implementation are to be found later in the text. In the last part the suggested methods of blood vessel detection are tested and the results evaluated.
Acceleration of Face Recognition Algorithm with Neural Compute Stick 2
Mičánková, Eva ; Beran, Jan (referee) ; Goldmann, Tomáš (advisor)
This thesis focuses on the issue of facial recognition in a face image using neural networks and its acceleration. It provides an overview of previously used techniques and addresses the use of currently dominant convolutional neural networks to solve this issue. The work also focuses on acceleration mechanisms that can be used in this area. Based on the knowledge of the issue, a system based on the concept of edge computing was created, which can be used as a home security system connected to an IP camera, which sends a notification about the presence of an unknown person in a guarded area.
Fake Face Detection in the Digital Images
Lapšanský, Tomáš ; Goldmann, Tomáš (referee) ; Orság, Filip (advisor)
V posledných rokoch môžeme pozorovať nárast a rýchly rozvoj neurónových sietí a umelej inteligencie v informačných technológiách, medzi ktoré patria aj deepfake fotografie a videá. Generatívne adverzné neurónové siete (GAN) sú toho jasným príkladom. V súčasnosti dokážu dosiahnuť výsledky, ktoré bežný človek nedokáže rozoznať od reality. Keďže tieto siete sa preto dajú zneužiť na rôzne účely, je potrebné vedieť rozlíšiť, čo je vygenerované a čo je skutočné. Táto práca skúma súčasné najmodernejšie riešenia neurónových sietí, ktoré môžu slúžiť ako vhodné modely na detekciu deepfake. Skúmame jednotlivé architektúry, ktoré sú vhodné ako základný model na detekciu, zaoberáme sa možnými vylepšeniami tohto modelu a vyvíjame niekoľko nových architektúr. Tie potom skúmame a vyhodnocujeme ich výsledky. V závere uvádzame diskusiu o výsledkoch a otvárame ďalšie otázky týkajúce sa tejto zložitej problematiky.
Face Recognition of Persons in Non-Frontal Positions
Horvát, Jozef ; Orság, Filip (referee) ; Goldmann, Tomáš (advisor)
This thesis deals with the creation of a model capable of recognizing people from faces in~images in non-frontal poses. The solution builds on the ArcFace algorithm and uses annotations of tilt and rotation angles of the face.
Security Implications of Deepfakes in Face Authentication
Šalko, Milan ; Goldmann, Tomáš (referee) ; Firc, Anton (advisor)
Deepfakes, médiá generované hlbokým strojovým učením, ktoré sú pre človeka nerozoznateľné od skutočných, zažívajú v posledných rokoch obrovský rozmach. O ich schopnosti oklamať ľudí už bolo napísaných niekoľko desiatok článkov. Rovnako závažný, ak nie závažnejší, môže byť problém, do akej miery sú voči nim zraniteľné systémy rozpoznávania tváre a hlasu. Zneužitie deepfakes proti automatizovaným systémom rozpoznávania tváre môže ohroziť mnohé oblasti nášho života, napríklad financie a prístup do budov. Táto téma je v podstate nepreskúmaným problémom. Cieľom tejto práce je preskúmať technickú realizovateľnosť útoku na rozpoznávanie tváre. Experimenty opísané v práci ukazujú, že tento útok je nielen uskutočniteľný, ale navyše útočník naň nepotrebuje veľa prostriedkov. V práci je opísaný aj rozsah tohto problému. V závere je opísaných aj niekoľko navrhovaných riešení tohto problému, ktoré vôbec nemusia byť náročné na implementáciu.
Probe for Determining the Surface Condition of Small Diameter Barrels
Procházka, Adam ; Goldmann, Tomáš (referee) ; Malaník, Petr (advisor)
This thesis focuses on the analysis of methods for measuring surface profiles, as well as the design and implementation of a device for their measurement. The implemented device utilizes non-contact triangulation profilometry techniques to determine the surface profile by projecting a laser line onto the surface and capturing the projected shape at a non-zero observation angle. It alse collects information about the surface’s image texture under LED illumination. In addition to data on image texture and profile, data on the position of the captured images is also collected using a sensor measuring the device’s distance from the end of the barrel. The work also includes the design and implementation of custom control modules for the chosen LED and laser diode. The output from the device is presented in a user application. The device is designed to function in tubes with diameters from 40 millimeters.
Detection and Recognition of Persons in a Multi-Camera System
Martinček, Ľuboš ; Beran, Jan (referee) ; Goldmann, Tomáš (advisor)
This thesis deals with the detection and recognition of people in a multi-camera system. In this thesis I describe general camera systems and their development and methods used for detection and recognition of persons. Based on this information, in the second part of the project I describe the design and implementation of a multi-camera system that is able to detect and recognize people. This thesis implements a combination of a YOLO detector and an Omni-Scale neural network based feature extractor.
Acceleration of a Neural Network for Face Detection in Low Light Conditions
Orava, Vojtěch ; Malinka, Kamil (referee) ; Goldmann, Tomáš (advisor)
The goal of this thesis is to build neural network for face detection in low light conditions, accelerate this network and compare it with other existing networks. Detection problem is solved with convolution neural network (CNN), which is trained on WIDER FACE and DARK FACE datasets. This CNN is accelerated by device Intel Neural Compute Stick 2. This work also summarise existing approaches in face detection (classic and neural networks based) and compares this approaches to the new ones. New detectors are based on TensorFlow Object Detection API. The best new model has average precision 47.1 % on custom validation dataset (detector YOLOv7-face has 42.8 % average precision). Speed of detection and influence of acceleration and quantization were also measured. With quantization some models could speed-up 3 times. Within this work, a GUI application for models tests was also developed. It can detect faces with newly created models and with some existing approaches.

National Repository of Grey Literature : 84 records found   1 - 10nextend  jump to record:
Interested in being notified about new results for this query?
Subscribe to the RSS feed.