Procjena modela za otkrivanja objekata korištenjem matrice konfuzije te srednje prosječne preciznosti (mAP)

undergraduate thesis

• Title (english): Evaluating object detection models using confusion matrix and mean average precision (mAP)
• Author: Ana Njirić
• Mentor: Tomislav Stipančić (mentor)
• Granter: University of Zagreb; Faculty of Mechanical Engineering and Naval Architecture; Zagreb
• Defense date and country: 2021-09-29, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Mechanical Engineering

Abstract

Tema ovog rada je evaluacija modela konvolucijske neuronske mreže (CNN) na CIFAR-100 bazi slika. Model mreže je izrađen u Python programskom jeziku koristeći biblioteku TensorFlow uz razne dodatne biblioteke. Procjena modela je odrađena korištenjem matrice konfuzija te srednje prosječne preciznosti (mAP). U prvom dijelu teorijski će te biti upoznati s konvolucijski neuronskim mrežama i alatima za evaluaciju rada tih modela. Nakon toga slijedi dio u kojem će biti prikazano i objašnjeno kodiranje evaluacije u programu Python. Na kraju se prikazuju i analiziraju rezultati te navode moguća poboljšanja rada mreže.

Abstract (english)

The topic of this paper is the evaluation of the convolutional neural network model (CNN) on a CIFAR-100 image base. The network model is developed in Python programming language using TensorFlow library with various additional libraries. Model evaluation was done using confusion matrix and mean average precision (mAP). In the first part you will theoretically be informed about convolutional neural networks and tools for evaluating the operation of these models. This is followed by the part presenting and explaining the coding of the evaluation in Python program. Finally, the results are presented and analysed and possible improvements in network operation are indicated.

• Keywords: konvolucijska neuronska mreža; CIFAR-100; matrica konfuzija; srednja prosječna preciznost (mAP); Python
• Keywords (english): convolutional neural network; CIFAR-100; confusion matrix; mean average precision (mAP); Python
• Language: croatian
• URN:NBN: urn:nbn:hr:235:907886
• Study programme: Title: Mechanical Engineering; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka strojarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka strojarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2021-09-22 14:21:23