On-Road Object Detection Based on Deep Residual Networks

Chen, Kang; Zhao, Qi; Lin, Yaorong; Zhang, Mengjie

doi:10.1007/978-3-319-70136-3_60

Cited by 3 publications

(3 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using this dataset for pretraining and subsequently fine-tuning on MS COCO significantly enhances performance outcomes. Model Backbone GFLOPS↓/FPS ↑ #params↓ mAP @[0.5,0.95] ↑ Epochs↓ URL Faster RCNN-DC5 (NeurIPS2015) [24] ResNet50 320/16 166M 21.4 37 Link Faster RCNN-FPN (NeurIPS2015) [24] ResNet50 180/26 42M 24.2 37 Link Faster RCNN-FPN (NeurIPS2015) [24] ResNet101 246/20 60M 25.2 -Link RepPoints v2-DCN-MS (NeurIPS2020) [119] ResNeXt101 -/--34.5 * 24 Link FCOS (ICCV2019) [56] ResNet50 177/17 -26.2 36 Link CBNet V2-DCN(ATSS [120]) (TIP2022) [70] Res2Net101 -/-107M 35.7 * 20 Link CBNet V2-DCN(Cascade RCNN) (TIP2022) [70] Res2Net101 -/-146M 37.4 * 32 Link DETR (ECCV2020) [31] ResNet50 86/28 41M 20.5 500 Link DETR-DC5 (ECCV2020) [31] ResNet50 187/12 41M 22.5 500 Link DETR (ECCV2020) [31] ResNet101 52/20 60M 21.9 -Link DETR-DC5 (ECCV2020) [31] ResNet101 253/10 60M 23.7 -Link ViT-FRCNN (arXiv2020) [32] --/--17.8 --RelationNet++ (NeurIPS2020) [35] ResNeXt101 -/--32.8 * -Link RelationNet++-MS (NeurIPS2020) [35] ResNeXt101 [41] ResNeXt101 -/--34.4 * -Link Dynamic DETR (ICCV2021) [44] ResNet50 -/--28.6 * --Dynamic DETR-DCN (ICCV2021) [44] ResNeXt101 -/--30.3 * --TSP-FCOS (ICCV2021) [55] ResNet101 255/12 -27.7 36 Link TSP-RCNN (ICCV2021) [55] ResNet101 254/9 -29.9 96 Link Mask R-CNN (ICCV2021) [57] Conformer-S/16 457.7/-56.9M 28.7 12 Link Conditional DETR-DC5 (ICCV2021) [65] ResNet101 262/-63M 27.2 108 Link SOF-DETR (2022JVCIR) [95] ResNet50 -/--21.7 -Link DETR++ (arXiv2022) [91] ResNet50 -/--22.1 --TOLO-MS (NCA2022) [69] --/57 -24.1 --Anchor DETR-DC5 (AAAI2022) [46] ResNet101 -/--25.8 50 Link DESTR-DC5 (CVPR2022) [68] ResNet101 299/-88M 28.2 50 -Conditional DETR v2-DC5 (arXiv2022) [66] ResNet101 228/-65M 26.3 50 -Conditional DETR v2 (arXiv2022) [66] Hourglass48 521/-90M 32.1 50 -FP-DETR-IN (ICLR2022) [80] --/-36M 26.5 50 Link DAB-DETR-DC5 (arXiv2...…”

Section: Uav123 [109]: This Dataset Contains 123 Videos Acquired Withmentioning

confidence: 99%

Extended Expectation Maximization for Under-Fitted Models

Rekavandi¹,

SEGHOUANE²,

Boussaid³

et al. 2022

Preprint

View full text Add to dashboard Cite

<p>In this paper, we generalize the well-known Expectation Maximization (EM) algorithm using the α− divergence for Gaussian Mixture Model (GMM). This approach is used in robust subspace detection when the number of parameters is kept small to avoid overfitting and large estimation variances. The level of robustness can be tuned by the parameterα. When α→1, our method is equivalent to the standard EM approach and for α <1 the method is robust against potential outliers. Simulation results show that the method outperforms the standard EM when it comes to mismatches between noise models and their realizations. In addition, we use the proposed method to detect active brain areas using collected functional Magnetic Resonance Imaging (fMRI) data during task-related experiments.</p>

show abstract

Section: Uav123 [109]: This Dataset Contains 123 Videos Acquired Withmentioning

confidence: 99%

Extended Expectation Maximization for Under-Fitted Models

Rekavandi¹,

SEGHOUANE²,

Boussaid³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Faster region-based convolutional neural network (R-CNN) [19], as the name faster R-CNN refers is an extension of fast R-CNN, as well as the name, suggests faster R-CNN is faster than its previous fast R-CNN which emphasizes the strongness of the region proposal network (RPN). By the use of RPN which refers to a fully convolutional network that is responsible for generating proposals with different aspect ratios and various scales.…”

Section: Introductionmentioning

confidence: 99%

Dissecting of the two-stages object detection models architecture and performance

Bouraya,

Belangour

2024

Bulletin EEI

View full text Add to dashboard Cite

Artificial intelligence (AI) is the discipline focused on enabling computers to operate autonomously without explicit programming. Within AI, computer vision is an emerging field tasked with endowing machines with the ability to interpret visual data from images and videos. Over recent decades, computer vision has found applications in diverse fields such as autonomous vehicles, information retrieval, surveillance, and understanding human behavior. Object detection, a key aspect of computer vision, employs deep neural networks to continually advance detection accuracy and speed. Its goal is to precisely identify objects within images or videos and assign them to specific classes. Object detection models typically consist of three components: a backbone network for feature extraction, a neck model for feature aggregation, and a head for prediction. The focus of this study lies on two stage detectors. This study aims to provide a comprehensive review of two stage detectors in object detection, followed by benchmarking to offer insights for researchers and scientists. By analyzing and understanding the efficacy of these models, this research seeks to guide future developments in the field of object detection within computer vision.

show abstract

“…Region-based CNN (R-CNN) [11] proposed a bounding-box regressionbased approach, later developed into Fast R-CNN [12], Faster R-CNN [13], and R-FCN (Region-based Fully Convolutional Network) [14]. Faster R-CNN uses the region proposal network (RPN) method to classify bounding boxes, followed by finetuning to process the bounding boxes [15,16]. One significant drawback of the two-stage architecture is its slow speed detection, resulting in the inability to produce real-time results as required for AV applications.…”

Section: Introductionmentioning

confidence: 99%

Fused Thermal and RGB Imagery for Robust Detection and Classification of Dynamic Objects in Mixed Datasets via Pre-Trained High-Level CNN

Ben-Shoushan

Brook

2023

Remote Sensing

View full text Add to dashboard Cite

Smart vehicles with embedded Autonomous Vehicle (AV) technologies are currently equipped with different types of mounted sensors, aiming to ensure safe movement for both passengers and other road users. The sensors’ ability to capture and gather data to be synchronically interpreted by neural networks for a clear understanding of the surroundings is influenced by lighting conditions, such as natural lighting levels, artificial lighting effects, time of day, and various weather conditions, such as rain, fog, haze, and extreme temperatures. Such changing environmental conditions are also known as complex environments. In addition, the appearance of other road users is varied and relative to the vehicle’s perspective; thus, the identification of features in a complex background is still a challenge. This paper presents a pre-processing method using multi-sensorial RGB and thermal camera data. The aim is to handle issues arising from the combined inputs of multiple sensors, such as data registration and value unification. Foreground refinement, followed by a novel statistical anomaly-based feature extraction prior to image fusion, is presented. The results met the AV challenges in CNN’s classification. The reduction of the collected data and its variation level was achieved. The unified physical value contributed to the robustness of input data, providing a better perception of the surroundings under varied environmental conditions in mixed datasets for day and night images. The method presented uses fused images, robustly enriched with texture and feature depth and reduced dependency on lighting or environmental conditions, as an input for a CNN. The CNN was capable of extracting and classifying dynamic objects as vehicles and pedestrians from the complex background in both daylight and nightlight images.

show abstract

On-Road Object Detection Based on Deep Residual Networks

Cited by 3 publications

References 8 publications

Extended Expectation Maximization for Under-Fitted Models

Extended Expectation Maximization for Under-Fitted Models

Dissecting of the two-stages object detection models architecture and performance

Fused Thermal and RGB Imagery for Robust Detection and Classification of Dynamic Objects in Mixed Datasets via Pre-Trained High-Level CNN

Contact Info

Product

Resources

About