Yazar "Arserim, M. Ali" seçeneğine göre listele

Listeleniyor 1 - 2 / 2
  • Küçük Resim
    Öğe
    Human posture prediction by deep learning
    (Dicle Üniversitesi Mühendislik Fakültesi, 2021) Kanpak, Hediye Nupelda; Arserim, M. Ali
    İnterpreting the human posture in human videos and pictures constitutes the most basic structure of human posture prediction. A system is created that decides what the movement is and what purpose it is made by evaluating pictures and videos. In this way, a structure has been created that determines and classifies human movements as an automatic system. A mechanism of motional meaning contained in the created system has been recognized in such away that the pattern is expressed. It is intended to take advantage of these components by taking instant information. A result was obtained by primarily inferring instant still images and eliminating time intervals that do not contain information range. A classification was made according to their accuracy. Based on the location coordinates of the images and videos, it was tried to determine how people might react in the neck stage. Thanks to the analysis performed through the joints with optical flow calculation, motion information was obtained and classifications and analyses expressing the power of motion were created. Motion information on the region determined in the image is determined by the detection of joints, revealing the power generated by movement. The created histograms provide ease of classification of motion. Based on the reliability of the descriptions, which include the concept of the time in a sequential way with the detection of joints, it was desired to create a sliding classification mechanism within the framework of these joints. As a result of this study, it was aimed to obtain a functional structure that can recognize and understand the autonomous movement of stationary or moving beings. An efficient structure has been created in terms of providing a useful and facilitating mechanism by solving the problems in estimation.
  • [ X ]
    Öğe
    Object Detection on FPGAs and GPUs by Using Accelerated Deep Learning
    (Ieee, 2019) Cambay, V. Yusuf; Ucar, Aysegul; Arserim, M. Ali
    Object detection and recognition is one of the main tasks in many areas such as autonomous unmanned ground vehicles, robotic and medical image processing. Recently, deep learning has been used by many researchers in these areas when the data measure is large. In particular, one of the most up-to-date structures of deep learning, Convolutional Neural Networks (CNNs) has achieved great success in this field. Real-time works related to CNNs are carried out by using GPU-Graphics Processing Units. Although GPUs provides high stability, they requires high power, energy consumption, and large computational load problems. In order to overcome this problem, it has started to used the Field Programmable Gate Arrays (FPGAs). In this article, object detection and recognition procedures were performed using the ZYNQ XC7Z020 development board including both the ARM processor and the FPGA. Real-time object recognition has been made with the Movidius USB-GPU externally plugged into the FPGA. The results are given with figures.