Sökning: "fine-grained classification problem"

Hittade 2 uppsatser innehållade orden fine-grained classification problem.

  1. 1. An Investigation of Low-Rank Decomposition for Increasing Inference Speed in Deep Neural Networks With Limited Training Data

    Uppsats för yrkesexamina på avancerad nivå, KTH/Skolan för elektroteknik och datavetenskap (EECS)

    Författare :Victor Wikén; [2018]
    Nyckelord :deep neural networks; convolutional neural networks; AlexNet; inference speed; optimization; low-rank tensor decomposition; fine-grained classification problem; dog breed classification; transfer learning;

    Sammanfattning : In this study, to increase inference speed of convolutional neural networks, the optimization technique low-rank tensor decomposition has been implemented and applied to AlexNet which had been trained to classify dog breeds. Due to a small training set, transfer learning was used in order to be able to classify dog breeds. LÄS MER

  2. 2. Deep Convolutional Neural Network for Effective Image Analysis : DESIGN AND IMPLEMENTATION OF A DEEP PIXEL-WISE SEGMENTATION ARCHITECTURE

    Master-uppsats, KTH/Skolan för informations- och kommunikationsteknik (ICT)

    Författare :Marco Ros Marti; [2017]
    Nyckelord :CNN; Co-CNN; segmentation; localization; RoI; masking; RoI masking; multi-task network; pixel resolution; overfitting; CNN; Co-CNN; segmentering; lokalisering; RoI; maskering; RoI- maskering; flera uppgiftsnätverk; pixelupplösning; övermontering; CNN; segmentació; localització; RoI; emmascarar; filtratge de RoIs; xarxa multi-tasca; resolució de píxel; overfitting;

    Sammanfattning : This master thesis presents the process of designing and implementing a CNN-based architecture for image recognition included in a larger project in the field of fashion recommendation with deep learning. Concretely, the presented network aims to perform localization and segmentation tasks. LÄS MER