Analiza značajki glasanja dupina korištenjem toplinske karte

undergraduate thesis

• Title (english): Understanding the features of dolphins whistles through heat maps
• Author: Jurica Jerinić
• Mentor: Nikola Mišković (mentor)
• Committee member: Nikola Mišković (član povjerenstva)
• Committee member: Roee Diamant (član povjerenstva)
• Committee member: Đula Nađ (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-09-06, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Ovaj završni rad fokusira se na analizu značajki glasanja dupina korištenjem toplinskih karata. Cilj rada je istražiti i primijeniti toplinske karte za vizualizaciju i detekciju zvižduka dupina putem neuronskih mreža, posebno VGG16 modela. Spektrogrami zvižduka tretirani su kao RGB slike, omogućujući precizno prepoznavanje strukture glasanja. Rad detaljno opisuje proces prikupljanja i obrade akustičnih podataka, razvoj i obuku modela, te generiranje i interpretaciju Grad-CAM toplinskih karata. Rezultati predstavljaju značajan korak prema unapređenju detekcije i razumijevanja glasanja dupina, pružajući temelj za daljnja istraživanja

Abstract (english)

This thesis focuses on the analysis of dolphin vocalization features using heat maps. The aim of the study is to explore and apply heat maps for the visualization and detection of dolphin whistles through neural networks, specifically the VGG16 model. Whistle spectrograms are treated as RGB images, allowing for precise recognition of vocalization structures. The thesis details the process of collecting and processing acoustic data, developing and training the model, and generating and interpreting Grad-CAM heat maps. The results represent a significant step towards improving the detection and understanding of dolphin vocalizations, providing a foundation for further research.

• Keywords: Toplinske karte; Grad-CAM; Duboko učenje; Neuronske mreže; Akustika; VGG16 Model; Detekcija zvižduka dupina
• Keywords (english): Heat Maps; Grad-CAM; Deep Learning; Neural Networks; Acoustics; VGG16 Model; Dolphin Whistle Detection
• Language: croatian
• URN:NBN: urn:nbn:hr:168:982942
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2024-10-12 20:56:41