Detekcija i raspoznavanje igraćih karata u slikama

undergraduate thesis

• Title (english): Detection and Recognition of Playing Cards in Images
• Author: Tin Plavec
• Mentor: Tomislav Hrkać (mentor)
• Committee member: Tomislav Hrkać (predsjednik povjerenstva)
• Committee member: Zorana Doždor (član povjerenstva)
• Committee member: Vjeko Kužina (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2022-07-04, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Prepoznavanje objekata danas se primjenjuje na razne način. Ovaj rad bavi se prepoznavanjem igraćih karata za trešetu. Korišten je vlastiti skup podataka od dvjestotinjak slika. Na svakoj slici nalazi se jedna ili više karata uz mogućnost preklapanja. Za bazni model odabran je SSD ResNet zbog svoje brzine i točnosti. Stoga je moguće istrenirani model koristiti za prepoznavanje u stvarnom vremenu. Treniranje je obavljeno koristeći Object Detection API iz open-source biblioteke TensorFlow. Istrenirani model gotovo savršeno prepoznaje slike s jednom kartom i s više karata bez preklapanja. Međutim, točnost pada na primjerima s preklapanjem karata, ali je i dalje zadovoljavajuća.

Abstract (english)

Object detection is widely used today. This undergraduate thesis is about detecting playing cards for tressette, a traditional Italian game. Own dataset has been made with around two hundred images. Every image has one or more cards that can overlap. SSD ResNet model is used as base because of its speed and accuracy. Therefore, the trained model can be used for real-time detection. The model is trained with Object Detection API from the open-source Tensor library. Images with one card or few cards without overlapping are recognized almost perfectly. However, overlapping cards have smaller accuracy, although acceptable.

• Keywords: detekcija objekata; strojno učenje; neuronske mreže; tensorflow; ssd resnet; trešeta; triestine
• Keywords (english): object detection; machine learning; neural networks; tensorflow; ssd resnet; tressette; triestine
• Language: croatian
• URN:NBN: urn:nbn:hr:168:463446
• 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: 2023-01-15 11:04:40