Searching Image Collections Using Deep Representations of Local Regions
Thesis title in Czech: | Vyhledávání v obrázkových kolekcích na základě lokálních regionů a reprezentací z hlubokých neuronových sítí |
---|---|
Thesis title in English: | Searching Image Collections Using Deep Representations of Local Regions |
Key words: | Hledání známeho objektu, Konvoluční neuronové sítě, Vyhledávání obrázků na základě obsahu, Explorace multimédií |
English key words: | Known-item search, Convolutional neural network, Content-based image retrieval, Multimedia exploration |
Academic year of topic announcement: | 2019/2020 |
Thesis type: | diploma thesis |
Thesis language: | angličtina |
Department: | Department of Software Engineering (32-KSI) |
Supervisor: | doc. RNDr. Jakub Lokoč, Ph.D. |
Author: | hidden - assigned and confirmed by the Study Dept. |
Date of registration: | 27.02.2020 |
Date of assignment: | 27.02.2020 |
Confirmed by Study dept. on: | 16.03.2020 |
Date and time of defence: | 14.09.2020 09:00 |
Date of electronic submission: | 30.07.2020 |
Date of submission of printed version: | 30.07.2020 |
Date of proceeded defence: | 14.09.2020 |
Opponents: | RNDr. Jiří Fink, Ph.D. |
Guidelines |
The thesis will investigate two particular approaches for searching large video datasets. The first approach will try to retrieve a searched scene based on a query collage consisting of example images organized on a canvas. The author will investigate various fusion methods, taking into account also positioning of the query images in the canvas. The second approach will be designed to search for a scene based on faces available in the dataset. The method will be based on visual exploration of the faces organized with respect to their similarities. In exploration approaches, the user can choose a similar face to the searched one, and continuously browse towards the searched face. The author will implement a basic interactive search prototype and evaluate search effectiveness with sets of experiments to estimate, how successfully users can retrieve particular scenes. |
References |
Kai Uwe Barthel and Nico Hezel and Konstantin Schall and Klaus Jung. Real-Time Visual Navigation in Huge Image Sets Using Similarity Graphs. arXiv:1910.06005
Jakub Lokoč, Gregor Kovalčík, Bernd Münzer, Klaus Schöffmann, Werner Bailer, Ralph Gasser, Stefanos Vrochidis, Phuong Anh Nguyen, Sitapa Rujikietgumjorn, and Kai Uwe Barthel. 2019. Interactive Search or Sequential Browsing? A Detailed Analysis of the Video Browser Showdown 2018. ACM Trans. Multimedia Comput. Commun. Appl. 15, 1, Article 29 (February 2019), 26 pages. Jakub Lokoč, Gregor Kovalčík, Tomáš Souček, Jaroslav Moravec, Přemysl Čech. 2019. VIRET: A video retrieval tool for interactive known-item search. In International Conference on Multimedia Retrieval (ICMR ’19), June 10–13, 2019, Ottawa, ON, Canada. ACM, New York, NY, USA, 5 pages DOBRANSKÝ, Marek. Object detection for video surveillance using the SSD approach. Praha, 2019. Diplomová práce. |