Pedestrian Detection Using R-CNN Object Detector

Masita, Katleho L; Hasan, Ali N.; Paul, Satyakama

doi:10.1109/la-cci.2018.8625210

Cited by 20 publications

(8 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since then, much effort has been made on improving the performance of pedestrian detection (Zhang et al, 2017;Pang et al, 2019;, and some others (Tian et al, 2015;Zhou and Yuan, 2018;Pang et al, 2019;Xie et al, 2020) target at handling occlusion. The work of Tian et al (2015), Masita et al (2018), Wei and Kehtarnavaz (2019), and Zhang et al (2020) proposed to detect pedestrians by addressing the fine-tuning problem, while Jiang et al (2016) and focus on speed up two-stage pedestrian detection algorithm. Since Fast R-CNN and Faster R-CNN have limited success for detecting small pedestrians due to the low resolution of their convolutional features (Zhang et al, 2016), Hu et al (2017) introduced semantic labels to improve pedestrians detections.…”

Section: Related Workmentioning

confidence: 99%

“…For instance, the work of Zeng et al (2014) proposes a transfer learning method for pedestrian detection that combines the ability to extract features from CNN and the similar ability to transfer features from the auto-encoder network. Masita et al (2018) use R-CNN via transfer learning to detect pedestrians. Initially, they fine-tune a pre-trained Alexnet CNN in the Penn-Fudan (170 images) and KTH Multiview Football (771 images) datasets.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Fine-Tuning Deep Learning Models for Pedestrian Detection

2021

View full text Add to dashboard Cite

Object detection in high resolution images is a new challenge that the remote sensing community is facing thanks to introduction of unmanned aerial vehicles and monitoring cameras. One of the interests is to detect and trace persons in the images. Different from general objects, pedestrians can have different poses and are undergoing constant morphological changes while moving, this task needs an intelligent solution. Fine-tuning has woken up great interest among researchers due to its relevance for retraining convolutional networks for many and interesting applications. For object classification, detection, and segmentation fine-tuned models have shown state-of-the-art performance. In the present work, we evaluate the performance of fine-tuned models with a variation of training data by comparing Faster Region-based Convolutional Neural Network (Faster R-CNN) Inception v2, Single Shot MultiBox Detector (SSD) Inception v2, and SSD Mobilenet v2. To achieve the goal, the effect of varying training data on performance metrics such as accuracy, precision, F1-score, and recall are taken into account. After testing the detectors, it was identified that the precision and recall are more sensitive on the variation of the amount of training data. Under five variation of the amount of training data, we observe that the proportion of 60%-80% consistently achieve highly comparable performance, whereas in all variation of training data Faster R-CNN Inception v2 outperforms SSD Inception v2 and SSD Mobilenet v2 in evaluated metrics, but the SSD converges relatively quickly during the training phase. Overall, partitioning 80% of total data for fine-tuning trained models produces efficient detectors even with only 700 data samples.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Fine-Tuning Deep Learning Models for Pedestrian Detection

2021

View full text Add to dashboard Cite

show abstract

“…Deep models for individual identification centres include learning [8], logical data learning, and handling of impediments. Forms of deep learning object identification [9] [10]can now mainly be isolated into two families: I two-stage locators, such as R-CNN [10], Fast R-CNN [11] and Faster R-CNN [12] [13] and their variations; and (ii) one-stage locators, like YOLO [12] and SSD. In two-stage locator recognition, the processed proposal is performed in steps, in the main stage, and organised into object classifications in the second stage.…”

Section: Related Workmentioning

confidence: 99%

A Novel Covid Prevention Method Using Deep Learning

Kaviya¹,

Vedhashree²,

Rani³

2020

2020 International Conference on Power, Energy, Control and Transmission Systems (ICPECTS)

View full text Add to dashboard Cite

World Health Organization has declared the information that the pandemic COVID-19 has harshly affected the world and has now contaminated in excess of 8,000,000 individuals around the world. Using face mask and social distancing serves as anupgraded security convention and overcomes the spread of the infection. In order to establish a covid free environment, the system implements an effective PC vision based methodology with a zeroed in on the constant robotized checking of individuals to identify both safe social separating and face masks. The method is openly put on by actualizing the model on embedded system based kit to screen action and recognize infringement using image processing. In this proposed framework present day profound learning calculations have been blended in with mathematical procedures for building a vigorous modular which covers three parts of discovery, following, and approval. Subsequently, the proposed framework favours the general public by sparing time and aides in bringing down the spread of Covid. It tends to be actualized successfully in current circumstance when lockdown is facilitated to assess people in open get-togethers, shopping centres; and so on computerized assessment lessens labour to examine general society and furthermore can be utilized in any spot

show abstract

“…Additionally, CNN’s are utilized in the field of the combination of spatial representation and time-series structure, i.e. moving object detection or video classification [Masita et.al, 2018; Li, 2017]. CNN’s provide significant performance enhancement minimizing the error rates of competing techniques in ImageNet competition 2012 [Krizhevsky et.al, 2017].…”

Section: Introductionmentioning

confidence: 99%

Development of LSTM&CNN Based Hybrid Deep Learning Model to Classify Motor Imagery Tasks

Uyulan

2020

Preprint

View full text Add to dashboard Cite

Recent studies underline the contribution of brain-computer interface (BCI) applications to the enhancement process of the life quality of physically impaired subjects. In this context, to design an effective stroke rehabilitation or assistance system, the classification of motor imagery (MI) tasks are performed through deep learning (DL) algorithms. Although the utilization of DL in the BCI field remains relatively premature as compared to the fields related to natural language processing, object detection, etc., DL has proven its effectiveness in carrying out this task.In this paper, a hybrid method, which fuses the one-dimensional convolutional neural network (1D CNN) with the long short-term memory (LSTM), was performed for classifying four different MI tasks, i.e. left hand, right hand, tongue, and feet movements. The time representation of MI tasks is extracted through the hybrid deep learning model training after principal component analysis (PCA)-based artefact removal process. The performance criteria given in the BCI Competition IV dataset A are estimated. 10-folded Cross-validation (CV) results show that the proposed method outperforms in classifying electroencephalogram (EEG)-electrooculogram (EOG) combined motor imagery tasks compared to the state of art methods and is robust against data variations. The CNN-LSTM classification model reached 95.62 % (±1.2290742) accuracy and 0.9462 (±0.01216265) kappa value for datasets with four MI-based class validated using 10-fold CV. Also, the receiver operator characteristic (ROC) curve, the area under the ROC curve (AUC) score, and confusion matrix are evaluated for further interpretations.

show abstract

Pedestrian Detection Using R-CNN Object Detector

Cited by 20 publications

References 21 publications

Fine-Tuning Deep Learning Models for Pedestrian Detection

Fine-Tuning Deep Learning Models for Pedestrian Detection

A Novel Covid Prevention Method Using Deep Learning

Development of LSTM&CNN Based Hybrid Deep Learning Model to Classify Motor Imagery Tasks

Contact Info

Product

Resources

About