Corpus ID: 10011895

Comparing Deep Neural Networks and Traditional Vision Algorithms in Mobile Robotics

@inproceedings{Lee2016ComparingDN,
  title={Comparing Deep Neural Networks and Traditional Vision Algorithms in Mobile Robotics},
  author={Andy Jongsuk Lee},
  year={2016}
}
We consider the problem of object detection on a mobile robot by comparing and contrasting two types of algorithms for computer vision. The first approach is coined ”traditional computer vision” and refers to using commonly known feature descriptors (SIFT, SURF, BRIEF, etc.) for object detection. The second approach uses Deep Neural Networks for object detection. We show that deep neural networks perform better than traditional algorithms, but discuss major trade offs surrounding performance… Expand

Figures and Tables from this paper

Deep Learning Object Detector Using a Combination of Convolutional Neural Network (CNN) Architecture (MiniVGGNet) and Classic Object Detection Algorithm
TLDR
Experimental results show that there are reasonable and efficient to combine classic object detection method with a deep learning classification approach, and the performance of this method can work in some specific use cases and effectively solving the problem of the inaccurate classification and detection of typical features. Expand
ROAD RECOGNITION USING FULLY CONVOLUTIONAL NEURAL NETWORKS
In the past few years an increasing tendency is perceptible in the field of research and application of neural networks and deep learning. For image classification and segmentation convolutionalExpand
Fast Object Recognition for Humanoid Robots by Using Deep Learning Models with Small Structure
TLDR
This study aims for Robotis-Op3 humanoid robot to recognize the different shaped objects with deep learning methods and proposes new models with small structure of Convolutional Neural Networks (CNNs) that produced higher performance than complex models. Expand
End-To-End Learning from Demonstation for Object Manipulation of Robotis-Op3 Humanoid Robot
TLDR
This study aimed for Robotis-Op3 humanoid robot to grasp the objects by learning from demonstrations based on vision by proposed a new algorithm using Convolutional Neural Networks and Long Short-Term Memory Networks. Expand
ANALYTICAL STUDY OF IMAGE CLASSSIFICATION USING DEEP LEARNING
TLDR
The main purpose of the work presented in this paper is the rise and development of machine learning, deep learning, CNN and to give an overview on using machine learning for image classification. Expand
Object Recognition in Remote Sensing Images Based on Modified Backpropagation Neural Network
TLDR
A modified backpropagation neural network is proposed to detect the objects in images and has better precision and robustness compared to the current approaches including the most conventional methods and certain profound learning methods to scene distinguish high-resolution remote sensing pictures. Expand
Android Smartphone Based Visual Object Recognition for Visually Impaired Using Deep Learning
TLDR
This paper proposes a deep learning based visual object recognition model, to guide a visually impaired person in an outdoor environment using a smartphone equipped with internet connectivity, and provides measurable high accuracy for recognition of 11 outdoor objects. Expand
Object Detection and their Localization for Visually Impaired Users
TLDR
A system which is a user friendly human computer interface enhanced with computer vision technology which will be able to support the visually impaired to localize and pick up objects used in their daily life is described. Expand
Generalizable semi-supervised learning method to estimate mass from sparsely annotated images
TLDR
The deep neural network (DNN) succeeds in capturing the mass of sugarcane accurately and surpasses older volumetric-based methods, despite highly varying lighting and material colors in the images. Expand
Real-time Driver Drowsiness Detection using Deep Learning
Every year thousands of lives pass away worldwide due to vehicle accidents, and the main reason behind this is the drowsiness in drivers. A drowsiness detection system will help to reduce thisExpand
...
1
2
...

References

SHOWING 1-3 OF 3 REFERENCES
Using the Forest to See the Trees: A Graphical Model Relating Features, Objects, and Scenes
TLDR
This work presents a conditional random field for jointly solving the tasks of object detection and scene classification, and proposes to use the scene context as an extra source of (global) information, to help resolve local ambiguities. Expand
CNN Features Off-the-Shelf: An Astounding Baseline for Recognition
TLDR
A series of experiments conducted for different recognition tasks using the publicly available code and model of the OverFeat network which was trained to perform object classification on ILSVRC13 suggest that features obtained from deep learning with convolutional nets should be the primary candidate in most visual recognition tasks. Expand
ImageNet classification with deep convolutional neural networks
TLDR
A large, deep convolutional neural network was trained to classify the 1.2 million high-resolution images in the ImageNet LSVRC-2010 contest into the 1000 different classes and employed a recently developed regularization method called "dropout" that proved to be very effective. Expand