Novel scene recognition using traindetector
The result's identifiers
Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F62690094%3A18450%2F19%3A50015952" target="_blank" >RIV/62690094:18450/19:50015952 - isvavai.cz</a>
Result on the web
<a href="http://dx.doi.org/10.1007/978-3-030-13469-3_59" target="_blank" >http://dx.doi.org/10.1007/978-3-030-13469-3_59</a>
DOI - Digital Object Identifier
<a href="http://dx.doi.org/10.1007/978-3-030-13469-3_59" target="_blank" >10.1007/978-3-030-13469-3_59</a>
Alternative languages
Result language
angličtina
Original language name
Novel scene recognition using traindetector
Original language description
Our ability to process the image keeps improving day by day, since the introduction of deep learning. Lastly, this contributed to the advance of object recognition through a Convolutional neural network and Place recognition, which is our concern in this paper. Through this research, it was observed a complexity in the extraction of the correct and relevant features for scene recognition. To address this issue, we extracted at the pixel level several subareas which contain more color intensity than other parts, and we went through each image once to build the feature representation of it. We also noticed that several available models based on Convolution Neural Network requires a Graphics Processing Units (GPU) for their implementation and are difficult to train. We propose in this paper, a novel Scene Recognition method using Single-Shot-Detector (SSD), Multi-modal Local-Receptive-Field (MM-LRF) and Extreme-Learning-Machine (ELM) that we named TrainDetector. It outperforms the state-of-the-art techniques when we apply it to three well-known scene recognition Datasets.
Czech name
—
Czech description
—
Classification
Type
D - Article in proceedings
CEP classification
—
OECD FORD branch
10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)
Result continuities
Project
—
Continuities
S - Specificky vyzkum na vysokych skolach
Others
Publication year
2019
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data specific for result type
Article name in the collection
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
ISBN
978-3-030-13468-6
ISSN
0302-9743
e-ISSN
—
Number of pages
9
Pages from-to
504-512
Publisher name
Springer Verlag
Place of publication
Berlin
Event location
Madrid
Event date
Nov 19, 2018
Type of event by nationality
WRD - Celosvětová akce
UT code for WoS article
—