Deep learning-based image recognition for autonomous driving

Fujiyoshi, Hironobu; Hirakawa, Tsubasa; Yamashita, Takayoshi

doi:10.1016/j.iatssr.2019.11.008

Cited by 351 publications

(160 citation statements)

References 5 publications

Supporting

Mentioning

156

Contrasting

Unclassified

Order By: Relevance

“…Agricultural farms are unconstrained natural environments or semi-constrained at very best. Machine learning has found intuitive applications in many fields, because of its adaptive learning ability, like in healthcare ( Ronneberger et al, 2015 ; Işın et al, 2016 ; Kauanova et al, 2017 ), autonomous driving ( Fujiyoshi et al, 2019 ; Hofmarcher et al, 2019 ; Imai, 2019 ), and weed and crop detection ( Grinblat et al, 2016 ; Mohanty et al, 2016 ; Dyrmann et al, 2017 ; Kussul et al, 2017 ; Fuentes et al, 2018 ). But very little work has been done in detecting fruits and classifying them according to their ripeness level.…”

Section: Related Workmentioning

confidence: 99%

DAM: Hierarchical Adaptive Feature Selection Using Convolution Encoder Decoder Network for Strawberry Segmentation

et al. 2021

View full text Add to dashboard Cite

Autonomous harvesters can be used for the timely cultivation of high-value crops such as strawberries, where the robots have the capability to identify ripe and unripe crops. However, the real-time segmentation of strawberries in an unbridled farming environment is a challenging task due to fruit occlusion by multiple trusses, stems, and leaves. In this work, we propose a possible solution by constructing a dynamic feature selection mechanism for convolutional neural networks (CNN). The proposed building block namely a dense attention module (DAM) controls the flow of information between the convolutional encoder and decoder. DAM enables hierarchical adaptive feature fusion by exploiting both inter-channel and intra-channel relationships and can be easily integrated into any existing CNN to obtain category-specific feature maps. We validate our attention module through extensive ablation experiments. In addition, a dataset is collected from different strawberry farms and divided into four classes corresponding to different maturity levels of fruits and one is devoted to background. Quantitative analysis of the proposed method showed a 4.1% and 2.32% increase in mean intersection over union, over existing state-of-the-art semantic segmentation models and other attention modules respectively, while simultaneously retaining a processing speed of 53 frames per second.

show abstract

Section: Related Workmentioning

confidence: 99%

DAM: Hierarchical Adaptive Feature Selection Using Convolution Encoder Decoder Network for Strawberry Segmentation

et al. 2021

View full text Add to dashboard Cite

show abstract

“…For example, in an autonomous car, there is enough energy to run one or more high-performance computing devices. In this case, the models must be accurate enough for specific tasks like pedestrian detection, so highly accurate DNN models are preferred [96].…”

Section: Edge-oriented Deep Neural Networkmentioning

confidence: 99%

Moving Deep Learning to the Edge

et al. 2020

View full text Add to dashboard Cite

Deep learning is now present in a wide range of services and applications, replacing and complementing other machine learning algorithms. Performing training and inference of deep neural networks using the cloud computing model is not viable for applications where low latency is required. Furthermore, the rapid proliferation of the Internet of Things will generate a large volume of data to be processed, which will soon overload the capacity of cloud servers. One solution is to process the data at the edge devices themselves, in order to alleviate cloud server workloads and improve latency. However, edge devices are less powerful than cloud servers, and many are subject to energy constraints. Hence, new resource and energy-oriented deep learning models are required, as well as new computing platforms. This paper reviews the main research directions for edge computing deep learning algorithms.

show abstract

“…CV concentrates on replicating parts of the complexity of the human visual system and enabling computers to identify and process objects in images and videos in the same way that humans do. Deep learning, especially convolutional neural networks (CNN), is vastly applied to the field of image recognition that can aid in autonomous driving; an overview is presented in Fujiyoshi et al [37]. The success of the ImageNet project has prompted many endeavors associated with self-driving technology using CV-real-time traffic light recognition [38], pedestrian detection [39]-just to name a few.…”

Section: The Smart City and Crimementioning

confidence: 99%

Enhancing City Sustainability through Smart Technologies: A Framework for Automatic Pre-Emptive Action to Promote Safety and Security Using Lighting and ICT-Based Surveillance

2020

View full text Add to dashboard Cite

The scope of the present paper is to promote social, cultural and environmental sustainability in cities by establishing a conceptual framework and the relationship amongst safety in urban public space (UPS), lighting and Information and Communication Technology (ICT)-based surveillance. This framework uses available technologies and tools, as these can be found in urban equipment such as lighting posts, to enhance security and safety in UPS, ensuring protection against attempted criminal activity. Through detailed literary research, publications on security and safety concerning crime and lighting can be divided into two periods, the first one pre-1994, and the second one from 2004–2008. Since then, a significant reduction in the number of publications dealing with lighting and crime is observed, while at the same time, the urban nightscape has been reshaped with the immersion of light-emitting diode (LED) technologies. Especially in the last decade, where most municipalities in the EU28 (European Union of all the member states from the accession of Croatia in 2013 to the withdrawal of the United Kingdom in 2020) are refurbishing their road lighting with LED technology and the consideration of smart networks and surveillance is under development, the use of lighting to deter possible attempted felonies in UPS is not addressed. To capitalize on the potential of lighting as a deterrent, this paper proposes a framework that uses existing technology, namely, dimmable LED light sources, presence sensors, security cameras, as well as emerging techniques such as artificial intelligence (AI)-enabled image recognition algorithms and big data analytics and presents a possible system that could be developed as a stand-alone product to alert possible dangerous situations, deter criminal activity and promote the perception of safety thus linking lighting and ICT-based surveillance towards safety and security in UPS.

show abstract

Deep learning-based image recognition for autonomous driving

Cited by 351 publications

References 5 publications

DAM: Hierarchical Adaptive Feature Selection Using Convolution Encoder Decoder Network for Strawberry Segmentation

DAM: Hierarchical Adaptive Feature Selection Using Convolution Encoder Decoder Network for Strawberry Segmentation

Moving Deep Learning to the Edge

Enhancing City Sustainability through Smart Technologies: A Framework for Automatic Pre-Emptive Action to Promote Safety and Security Using Lighting and ICT-Based Surveillance

Contact Info

Product

Resources

About