Scene Understanding Using Deep Learning

Husain, Farzad; Dellen, Babette; Torras, Carme

doi:10.1016/b978-0-12-811318-9.00020-x

Cited by 19 publications

(14 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In contrast to the convolutional layers, fully-connected layers comprise more parameters that lead to them being harder to train [38]. The last layer for prediction in the CNN architecture is the regression layer that makes the relationship between an ideal target and one or more independent features (the input data) [30,39].…”

Section: Convolutional Neural Networkmentioning

confidence: 99%

“…Training a CNN means that the model has to be learned the best weights and biases in each layer for minimizing a defined cost function. The procedure of the minimization of a cost function is executed repetitively, utilizing a gradient descent technique that includes the computation of fractional derivatives from the cost function [38].…”

Section: Convolutional Neural Networkmentioning

confidence: 99%

See 1 more Smart Citation

RecPOID: POI Recommendation with Friendship Aware and Deep CNN

Safavi

Jalali

2021

Future Internet

View full text Add to dashboard Cite

In location-based social networks (LBSNs), exploit several key features of points-of-interest (POIs) and users on precise POI recommendation be significant. In this work, a novel POI recommendation pipeline based on the convolutional neural network named RecPOID is proposed, which can recommend an accurate sequence of top-k POIs and considers only the effect of the most similar pattern friendship rather than all user’s friendship. We use the fuzzy c-mean clustering method to find the similarity. Temporal and spatial features of similar friends are fed to our Deep CNN model. The 10-layer convolutional neural network can predict longitude and latitude and the Id of the next proper locations; after that, based on the shortest time distance from a similar pattern’s friendship, select the smallest distance locations. The proposed structure uses six features, including user’s ID, month, day, hour, minute, and second of visiting time by each user as inputs. RecPOID based on two accessible LBSNs datasets is evaluated. Experimental outcomes illustrate considering most similar friendship could improve the accuracy of recommendations and the proposed RecPOID for POI recommendation outperforms state-of-the-art approaches.

show abstract

Section: Convolutional Neural Networkmentioning

confidence: 99%

Section: Convolutional Neural Networkmentioning

confidence: 99%

RecPOID: POI Recommendation with Friendship Aware and Deep CNN

Safavi

Jalali

2021

Future Internet

View full text Add to dashboard Cite

show abstract

“…In addition, deep learning has an active merit of transfer learning that a myriad of pretraining networks and public datasets have provided benefits for training numerous traffic scenes. For vehicle-related scenes, to understand the objects, scenes, and events in the video, the deep learning neural network attempts to emulate the high-level abstraction in the data and encode it as a robust representation (Husain, et al, 2017).…”

Section: Background and Motivationmentioning

confidence: 99%

Vehicle-Related Scene Understanding Using Deep Learning

Liu

Neuyen

Yan

2020

Communications in Computer and Information Science

View full text Add to dashboard Cite

Automated driving technology is an inevitable trend in the future development of transportation, it is also one of the eminent achievements in the matter of artificial intelligence. Primarily deep learning produces a significant contribution to the progression of automatic driving. Deep learning not only promotes autonomous vehicles to sense and identify the surrounding environment, but also identifies and classifies various information regarding to vehicles. With the upgrades and improvement of deep learning technology, it can be promptly and readily learned and employed. A large number of pretraining networks and public datasets have provided convenience for training numerous traffic scenes. Nevertheless, automated driving technology is not flexible enough to understand scenes in complex traffic environments, with regard to traffic rules and transportation facilities in various countries. There is no algorithm so far designed for all traffic scenes. In this thesis, our contributions are that we primarily deal with the issue of understanding of vehicle-related scene using deep learning. To the best of our knowledge, this is the first time that we utilize Auckland traffic environment as an analysis object for scene understanding. Moreover, automatic scene segmentation and object detection are coalesced for traffic scene understanding. The techniques based on deep learning dramatically decrease human manipulations. Furthermore, the datasets in this project provide a large amount of Auckland traffic data. Meanwhile, the performance of CNN processing is consolidated by combining with vehicle detection outcome.

show abstract

“…This thesis focuses on the most popular variation of feed-forward ANN called Convolutional Neural Network (CNN). CNN provides solutions to a range of problems, from computer vision [4][5][6][7][8][9][10][11], raw audio generation [12], to general data analytics [13], to name a few. Each network consists of a repeating set of layers (Convolution, Non-linear Activation, Pooling, Fully Connected) interconnected sequentially.…”

Section: Convolutional Neural Networkmentioning

confidence: 99%

“…Convolutional Neural Networks (CNN) provide solutions to a range of problems, from computer vision [4][5][6][7][8][9][10][11], raw audio generation [12], to general data analytics [13], to name some. Each network consists of a set of layers, where each layer transforms its input using a set of filters and produces a multi-channel output.…”

Section: Introductionmentioning

confidence: 99%

Techniques in enhancing computation and understanding of convolutional neural networks

Abdiyeva¹

View full text Add to dashboard Cite

show abstract

Scene Understanding Using Deep Learning

Cited by 19 publications

References 26 publications

RecPOID: POI Recommendation with Friendship Aware and Deep CNN

RecPOID: POI Recommendation with Friendship Aware and Deep CNN

Vehicle-Related Scene Understanding Using Deep Learning

Techniques in enhancing computation and understanding of convolutional neural networks

Contact Info

Product

Resources

About