The following information was submitted:
Transactions: INTERNATIONAL JOURNAL of CIRCUITS, SYSTEMS and SIGNAL PROCESSING
Transactions ID Number: 20-749
Full Name: Emel Arslan
Position: Doctor (Researcher)
Age: ON
Sex: Female
Address: Research and Application Center for Computer Sciences, Istanbul University, Istanbul, Beyazýt, 34452
Country: TURKEY
Tel: +90 212 4000093
Tel prefix:
Fax: :+90 212 4400094
E-mail address: earslan@istanbul.edu.tr
Other E-mails: emel@posta.cc
Title of the Paper: Feature extraction of objects in moving images and implementation of the purification algorithm on the CNN Universal Machine
Authors as they appear in the Paper: Emel Arslan, Zeynep Orman, Sabri Arik
Email addresses of all the authors: earslan@istanbul.edu.tr, ormanz@istanbul.edu.tr, ariks@istanbul.edu.tr
Number of paper pages: 8
Abstract: This paper presents an algorithm that detects certain objects in a moving image and extracts their features. We named this algorithm as the purification algorithm because it is also used for purifying the remains of detected objects and for saving each purified object as a separate image file. The algorithm is implemented on the Bi-i Cellular Vision System which is a Cellular Neural Network(CNN) Universal Machine. The CNN Universal Machine is known as the analogical array computer and it contains two processors which can work interactively with each other. These processors are the ACE16k that is the hardware implementation of CNNs and the Digital Signal Processor(DSP). The purification algorithm is implemented with two different applications. In the first application, all phases of the algorithm are implemented just on the DSP. In the second application, the morphological operations of the algorithm are performed on the ACE16k and all other operations are performed!
on the DSP. Therefore, in the latter one, the application is run in coordination with both the ACE16k processor and the DSP. The obtained results are evaluated in terms of the run-time of the purification algorithm to show the comparison of these applications. Experimental results show that the performance of the proposed algorithm is good.
Keywords: ACE16k, Bi-i Cellular Vision System, Cellular Neural Networks, CNN Universal Machine, Digital Signal Processor, Image processing
EXTENSION of the file: .pdf
Special (Invited) Session: Detection of objects in moving images and implementation of the purification algorithm on Analog CNN and DSP processors
Organizer of the Session: 656-129
How Did you learn about congress: Tolga Ensari (ensari@istanbul.edu.tr), Selcuk Sevgen (sevgens@istanbul.edu.tr), Ruya Samli (ruyasamli@istanbul.edu.tr)
IP ADDRESS: 78.181.99.49
