Sustav za detekciju pješaka temeljen na dubokom učenju

master's thesis

• Title (english): Pedestrian Detection System Based on Deep Learning
• Author: Filip Gobin
• Mentor: Zoran Kalafatić (mentor)
• Committee member: Zoran Kalafatić (predsjednik povjerenstva)
• Committee member: Marko Čupić (član povjerenstva)
• Committee member: Tomislav Hrkać (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2016-07-14, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

U ovom radu prikazana je detekcija pješaka temeljena na konvolucijskim neuronskim mrežama koristeći Faster R-CNN sustav. Faster R-CNN je prvo istreniran nad ImageNet bazom slika a zatim je fino ugađan nad slikama pješaka. U radu su testirane dvije baze pješaka, INRIA i Caltech te su prikazani i opisani rezultati detekcije.

Abstract (english)

This thesis deals with pedestrian detection in images based on convolutional neural networks using Faster R-CNN system. Faster R-CNN is trained on ImageNet image set and then fine-tuned on pedestrian images using two datasets: INRIA and Caltech. In the last part of work results and possible improvements are briefly discussed.

• Keywords: Konvolucijske neuronske mreže; R-CNN; Fast R-CNN; Faster R-CNN; Prepoznavanje pješaka; ImageNet; INRIA; Caltech; Računalni vid; Raspoznavanje uzoraka
• Keywords (english): Convolutional neural networks; R-CNN; Fast R-CNN; Faster R-CNN; pedestrian detection; ImageNet; INRIA; Caltech; computer vision; pattern recognition
• Language: croatian
• URN:NBN: urn:nbn:hr:168:976635
• Study programme: Title: Computing; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka računarstva (magistar/magistra inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2018-02-06 08:55:32