A Comparative Study of Different CNN Encoders for Monocular Depth Prediction

Aladem, Mohamed; Chennupati, Sumanth; Shair, Zaid El; Rawashdeh, Samir A.

doi:10.1109/naecon46414.2019.9057857

Cited by 4 publications

(3 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It provides mainly an evaluation of multiple monocular methods in indoor environments. Also, in [30] a comparison and evaluation of multiple encoder architectures for depth estimation is proposed. None of these papers went for comprehensive comparison and evaluation of both stereoscopic and monocular depth estimation in road environment, which is the aim of this paper.…”

Section: Evaluation Of Depth Estimation Methodsmentioning

confidence: 99%

A Comparative Study of Deep Learning-based Depth Estimation Approaches: Application to Smart Mobility

Mauri

Khemmar

Decoux

et al. 2021

2021 8th International Conference on Smart Computing and Communications (ICSCC)

View full text Add to dashboard Cite

Section: Evaluation Of Depth Estimation Methodsmentioning

confidence: 99%

A Comparative Study of Deep Learning-based Depth Estimation Approaches: Application to Smart Mobility

Mauri

Khemmar

Decoux

et al. 2021

2021 8th International Conference on Smart Computing and Communications (ICSCC)

View full text Add to dashboard Cite

“…The camera is calibrated and synchronized with the Velodyne LiDAR and with the GPS/IMU sensor [ 35 ]. KITTI includes 3D object tracking tags for different classes of objects (cars, trucks, streetcars, pedestrians, and cyclists) and serves as a baseline for training and evaluation of stereoscopy, optical flow [ 36 ], visual odometry, 2D and 3D object detection [ 2 , 3 , 4 , 5 , 6 , 7 ], depth estimation [ 8 , 9 , 10 , 11 , 12 , 13 , 14 ], and 2D tracking [ 37 , 38 , 39 ].…”

Section: Related Workmentioning

confidence: 99%

“…Our CNN approaches that use datasets are dedicated to real-time 3D object detection and tracking for both road and railway smart mobility. The following tasks require a dataset for training such as: 2D object detection [ 2 , 3 , 4 , 5 , 6 , 7 ], Object distance estimation [ 8 , 9 , 10 , 11 , 12 , 13 , 14 ], 3D object centers, 3D object detection (center, dimension, and orientation) [ 2 , 3 , 4 ], Object tracking [ 2 ], Semantic segmentation [ 5 , 15 , 16 , 17 ]. …”

Section: Introductionmentioning

confidence: 99%

Road and Railway Smart Mobility: A High-Definition Ground Truth Hybrid Dataset

Khemmar

Mauri

Dulompont

et al. 2022

Sensors

View full text Add to dashboard Cite

A robust visual understanding of complex urban environments using passive optical sensors is an onerous and essential task for autonomous navigation. The problem is heavily characterized by the quality of the available dataset and the number of instances it includes. Regardless of the benchmark results of perception algorithms, a model would only be reliable and capable of enhanced decision making if the dataset covers the exact domain of the end-use case. For this purpose, in order to improve the level of instances in datasets used for the training and validation of Autonomous Vehicles (AV), Advanced Driver Assistance Systems (ADAS), and autonomous driving, and to reduce the void due to the no-existence of any datasets in the context of railway smart mobility, we introduce our multimodal hybrid dataset called ESRORAD. ESRORAD is comprised of 34 videos, 2.7 k virtual images, and 100 k real images for both road and railway scenes collected in two Normandy towns, Rouen and Le Havre. All the images are annotated with 3D bounding boxes showing at least three different classes of persons, cars, and bicycles. Crucially, our dataset is the first of its kind with uncompromised efforts on being the best in terms of large volume, abundance in annotation, and diversity in scenes. Our escorting study provides an in-depth analysis of the dataset’s characteristics as well as a performance evaluation with various state-of-the-art models trained under other popular datasets, namely, KITTI and NUScenes. Some examples of image annotations and the prediction results of our 3D object detection lightweight algorithms are available in ESRORAD dataset. Finally, the dataset is available online. This repository consists of 52 datasets with their respective annotations performed.

show abstract