Object security using Edge AI
| Thesis title in Czech: | Zabezpečení objektů pomocí Edge AI |
|---|---|
| Thesis title in English: | Object security using Edge AI |
| Key words: | Edge AI|Detekce krádeže|Bezpečnost objektů|OpenCV AI Kit |
| English key words: | Edge AI|Theft detection|Object Security|OpenCV AI Kit |
| Academic year of topic announcement: | 2022/2023 |
| Thesis type: | Bachelor's thesis |
| Thesis language: | angličtina |
| Department: | Department of Theoretical Computer Science and Mathematical Logic (32-KTIML) |
| Supervisor: | RNDr. David Obdržálek, Ph.D. |
| Author: | hidden - assigned and confirmed by the Study Dept. |
| Date of registration: | 02.11.2022 |
| Date of assignment: | 02.11.2022 |
| Confirmed by Study dept. on: | 20.10.2023 |
| Guidelines |
| The student shall study possibilities of providing physical security for a chosen object using Edge-native Artificial Intelligence. The resulting theoretical findings are expected to be used for developing and implementing such system in a pilot implementation.
The work is expected to consist of both software and hardware, where the latter would encompass the sensors, computing platform, and, if needed, means of providing physical security on a demonstration level. The software part of the work should involve detecting specified danger to the object using AI techniques. |
| References |
| SICILIANO, Bruno, and KHATIB, Oussama. Springer handbook of robotics. Second edition. Cham: Springer, 2016. ISBN 9783319325507
FRADEN, Jacob. Handbook of Modern Sensors Physics, Designs, and Applications. 5th ed. 2016. Cham: Springer International Publishing, 2016. ISBN 1-78684-326-9 OSHIDA, Sota R. Computer vision. New York: Nova Science Publishers, 2011. ISBN 1-62081-701-2 ŠONKA, Milan, HLAVÁČ, Václav and BOYLE, Roger. Image processing, analysis, and machine vision. 4th ed. Stamford: Cengage Learning, 2015. ISBN 978-1-133-59360-7 ŠIKUDOVÁ, Elena et al.. Počítačové videnie: detekcia a rozpoznávanie objektov. Praha: Wikina, 2014. ISBN 978-80-87925-06-5 |
| Preliminary scope of work |
| S rostoucí popularitou internetového nakupování se zvyšuje počet krádeží balíčků. Tato práce se zabývá využitím kombinace umělé inteligence a edge computing - Edge AI, k zajištění fyzické bezpečnosti s hlavním zaměřením na ochranu balíčků před krádeží. Práce nejprve analyzuje současná řešení a poukazuje na jejich omezení. Následně představuje nový systém pro detekci krádeží, který využívá kamery OAK a kombinaci neuronových sítí. Tento systém kombinuje veřejně dostupné modely a vlastně trénovaný segmentační model, což zajišťuje efektivní detekční schopnosti. Účinnost navrhovaného řešení založeného na Edge AI je demonstrována prostřednictvím designu a implementace speciálně vyvinutého hardwaru. Tento hardware ilustruje funkčnost našeho systému skrze praktickou ukázku v reálných podmínkách. |
| Preliminary scope of work in English |
| As e-commerce continues to expand globally, package theft has emerged as a significant issue. This thesis explores the application of Edge-native Artificial Intelligence for providing physical security, focusing on safeguarding packages against theft. The thesis first examines current solutions, highlighting their limitations, and then proposes a novel theft detection system that utilizes OAK cameras to execute a combination of neural networks. The system incorporates both publicly available models and a custom-trained segmentation model, ensuring robust and effective detection capabilities. The effectiveness of the proposed Edge AI-based solution is demonstrated through the design and implementation of custom hardware, which serves to showcase the proof-of-concept system's functionality in real-world scenarios. |