14699

13230

كارشناسي ارشد

الكترونيك

اصفهان : دانشگاه صنعتي اصفهان

1398

چهارده، 101ص.: مصور، جدول، نمودار

مسعود سيدي

رسول دهقاني، اميررضا احمدي مهر

1398/04/05

كتابنامه

مهندسي برق

مهندسي برق و كامپيوتر

1398/04/05

2544078

Study and Design of Vision Chips Based on Processing at Block of Pixels Level Farzaneh Babashahi f babashahi@ec iut ac ir March 21 2015 Department of Electrical and Computer Engineering Isfahan University of Technology Isfahan 84156 83111 Iran Degree M Sc Language Farsi Supervisor Sayed Masoud Sayedi m sayedi@cc iut ac ir Abstract Vision chips have many applications in tracking systems object recognition image segmentation etc The advancementof CMOS technology and reduction of transistors sizes have made many progress in improving of image quality reductionof area and power consumption and increasing of computational speed By using this technology and employing high densitycircuits photo sensors and processing units are integrated in one chip Elimination of the need for data transmission betweenthe sensors and the off chip circuits reduces power consumption and increases the processing speed Vision chips based ontheir usage are designed in five different architectures These architectures are different in terms of computational speed power consumption and chip area Block based vision chip is one of these architectures that consists of some similar circuitblocks that each contains a group of photo pixels e g a 16 16 array pixels and one processor that works in parallel withother block processors The main advantage of block based architecture is its scalability in terms of resolution and imagesize In this work this architecture is implemented on FPGA by using VHDL codes and MRI images are processed in thisstructure Basic reference information is obtained by using normal MRI images and fuzzy clustering This information isused to process MRI images of patients with Alzheimer s disease In addition an algorithm is introduced to find each blockdistance from the center part of the brain By combining the result of the intensity fuzzy clustering algorithm with the spatialinformation of pixel lesion and its location are detected in patient MRI images A fixed point ALU is implemented andused in each block for fuzzy clustering computations The implemented ALU has an accuracy of three decimal places Theproper size of the blocks is obtained and based on that the clustering algorithm is implemented on FPGA Comparing theresults with the software results confirms the correct implementation of the algorithm Key Words Vision Chip Pixel Array Block Based Architecture MRI Images Fuzzy Clustering Alzheimer s Disease

مسعود سيدي

رسول دهقاني، اميررضا احمدي مهر