The Importance of Loss Functions for Increasing the Generalization Abilities of a Deep Learning-Based Next Frame Prediction Model for Traffic Scenes

The Importance of Loss Functions for Increasing the Generalization Abilities of a Deep Learning-Based Next Frame Prediction Model for Traffic Scenes

This paper analyzes in detail how different loss functions influence the generalization abilities of a deep learning-based next frame prediction model for traffic scenes. Our prediction model is a convolutional long-short term memory (ConvLSTM) network that generates the pixel values of the next fra...

詳細記述

書誌詳細
主要な著者: Sandra Aigner, Marco Körner
フォーマット: 論文
言語:English
出版事項: MDPI AG 2020-04-01
シリーズ:Machine Learning and Knowledge Extraction
主題:
traffic scene prediction
video prediction
generalization
convolutional LSTMs
recurrent neural networks
machine learning
オンライン･アクセス:https://www.mdpi.com/2504-4990/2/2/6

インターネット

https://www.mdpi.com/2504-4990/2/2/6

類似資料