Significantly improving human detection in low-resolution images by retraining YOLOv3

Pouyan, Shima; Charmi, Mostafa; Azarpeyvand, Ali; Hassanpoor, Hossein

doi:10.1109/csicc52343.2021.9420554

Cited by 2 publications

(2 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This division makes some grid cells and the position of object's center should fall into that grid cell to be labeled as detected object. Resolution of surveillance videos are usually low and head cover detection module should also overcome this challenge.The YOLO algorithms can be retrained by low-resolution images to make an accurate detection result [45]. YOLO v1, v2 and v3 can detect type of an object and its position at an image [46].…”

Section: ) Head Cover Detectionmentioning

confidence: 99%

Propounding First Artificial Intelligence Approach for Predicting Robbery Behavior Potential in an Indoor Security Camera

et al. 2023

Self Cite

View full text Add to dashboard Cite

Crime prediction in video-surveillance systems is required to prevent incident and protect assets. In this sense, our article proposes first artificial intelligence approach for Robbery Behavior Potential (RBP) prediction and detection in an indoor camera. Our method is based on three detection modules including head cover, crowd and loitering detection modules for timely actions and preventing robbery. The two first modules are implemented by retraining YOLOV5 model with our gathered dataset which is annotated manually. In addition, we innovate a novel definition for loitering detection module which is based on DeepSORT algorithm. A fuzzy inference machine renders an expert knowledge as rules and then makes final decision about predicted robbery potential. This is laborious due to: different manner of robber, different angle of surveillance camera and low resolution of video images. We accomplished our experiment on real world video surveillance images and reaching the F1-score of 0.537. Hence, to make an experimental comparison with the other related works, we define threshold value for RBP to evaluate video images as a robbery detection problem. Under this assumption, the experimental results show that the proposed method performs significantly better in detecting the robbery as compared to the robbery detection methods by distinctly report with F1-score of 0.607. We strongly believe that the application of the proposed method could cause reduction of robbery detriment in a control center of surveillance cameras by predicting and preventing incident of robbery. On the other hand, situational awareness of human operator enhances and more cameras can be managed.

show abstract

Section: ) Head Cover Detectionmentioning

confidence: 99%

Propounding First Artificial Intelligence Approach for Predicting Robbery Behavior Potential in an Indoor Security Camera

et al. 2023

Self Cite

View full text Add to dashboard Cite

show abstract

“…Images of this dataset consist of masked and no-masked human and this is also a challenging face detection dataset. The gathered dataset can also named as a collection with several constraint including: ultra low resolution, various face position and diverse occlusion which have been used in [55] and [56] for creating a robust detection framework. These images also split into 40%, 10% and 50% for training, validation and testing set respectively.…”

Section: Our Prepared Datasetmentioning

confidence: 99%

WT-BiFPN: A Novel Approach for Efficient Face Detection using Improved EfficientDet Architecture and Dual-Tree Complex Wavelet Transform

Pouyan

Charmi

Azarpeyvand

et al. 2023

Preprint

Self Cite

View full text Add to dashboard Cite

Face detection is a crucial task in computer vision and image processing with various practical applications including security, surveillance and entertainment. In recent years, approaches based on deep learning methods improved significantly with high-accuracy detection results. In this paper, we propose a new framework for anefficient face detection based on improved EfficientDet architecture and wavelet transform. Our method utilizes a combination of innovated bi-directional feature pyramid network (BiFPN) and a dual-tree complex wavelet transform called WT-BiFPN to ameliorate feature representation of faces at multiple scales. We evaluate our approach on two benchmark datasets, including WIDER FACE and our gathered dataset with specific details close to real-world images. Our proposed architecture achieves more than 5% performance improvement on previous state-of-the-art EfficientDet-based methods in terms of mean average precision (mAP). Our method provides an effective accurate face detection solution for several applications and can prevail over low resolution and occlusion in images.

show abstract

Significantly improving human detection in low-resolution images by retraining YOLOv3

Cited by 2 publications

References 22 publications

Propounding First Artificial Intelligence Approach for Predicting Robbery Behavior Potential in an Indoor Security Camera

Propounding First Artificial Intelligence Approach for Predicting Robbery Behavior Potential in an Indoor Security Camera

WT-BiFPN: A Novel Approach for Efficient Face Detection using Improved EfficientDet Architecture and Dual-Tree Complex Wavelet Transform

Contact Info

Product

Resources

About