The rapid growth of visual surveillance and personal identification systems in smart environments has empowered the development of real-time face recognition, even under imperfect conditions. Among these challenges, r...
详细信息
ISBN:
(纸本)9798350362770;9798350362763
The rapid growth of visual surveillance and personal identification systems in smart environments has empowered the development of real-time face recognition, even under imperfect conditions. Among these challenges, recognizing faces in blurry images stands as a significant barrier to achieving consistent identification results. In this paper, we combine advanced machine learning algorithms and deep learning frameworks, integrating Generative Adversarial Networks (GANs), specifically the Generative Facial Prior (GFPGAN), to restore clarity to blurred facial images. The primary objective is to enhance the accuracy of face recognition in real-world scenarios where image quality may be compromised. In the proposed approach, the GFPGAN model serves as the primary mechanism for blind face restoration, transforming blurred images into clear and recognizable visuals. Upon achieving clarity in the image, a customized Convolutional Neural Network (CNN) is deployed to perform face recognition tasks. To showcase the efficiency of the proposed system, a comprehensive comparative analysis is carried out, comparing the results of the pretrained models VGG-Face, FaceNet, CNN, and DeepFace. Our experiments showed a high validation accuracy of 83.72% for FaceNet. Our experimental results demonstrate that the combination of GFPGAN preprocessing and deep learning models significantly improves blurry images face recognition accuracy.
The task of video-based human recognition is complicated by many factors such as imaging distortions, imaging range, lack of frames, arbitrary pose, occlusions, air turbulence, and changing clothes. This work presents...
详细信息
ISBN:
(纸本)9798350337266
The task of video-based human recognition is complicated by many factors such as imaging distortions, imaging range, lack of frames, arbitrary pose, occlusions, air turbulence, and changing clothes. This work presents the first study that utilizes single-frame binary silhouettes and their auxiliary representations for human recognition under extreme distortions. The proposed representation is compact, modular, and robust to distortions, allowing for easy deployability for long-range recognition. Quantitative metrics are reported on long-range dataset such as Briar, demonstrating the robustness of the proposed approach to common challenges of video-based recognition in the wild. The proposed single-frame method is compared against gait techniques using limited frames, outperforming in most cases. Performance is also compared to grayscale images with varying ranges, environments, and changing clothes, where the proposed model outperforms grayscale images. Under consistent conditions, the proposed model still augments the performance of the baseline grayscale model by over 15%.
This paper introduces a screen control mechanism for long distance interaction between Human and computer. This Paper provides solution for controlling the screen elements of different applications and browsers throug...
详细信息
Facial expression recognition(FER) is one of the important research in computervision, which has been widely applied in human-computer interaction, education, healthcare, transportation, etc. However, the wide applic...
An intelligent fatigue-driving detection system was developed utilizing machine vision. In the system, image acquisition, processing, and fatigue feature extraction were integrated to enhance precision and operational...
详细信息
Facial recognition is becoming more and more in demand as technology advances. Among the several facial recognition algorithms that are available, the Linear Binary pattern Classifier and the Haar cascade is used Due ...
详细信息
In this paper, we introduce an Feature Aggregate Side Network (FASN), a simple, efficient, and easy-to-train method for open-vocabulary semantic segmentation. Building upon existing models based on the CLIP-Side Netwo...
详细信息
ISBN:
(纸本)9798350359329;9798350359312
In this paper, we introduce an Feature Aggregate Side Network (FASN), a simple, efficient, and easy-to-train method for open-vocabulary semantic segmentation. Building upon existing models based on the CLIP-Side Network framework, we address the issue of CLIP-generated features lacking pixel-level recognition capability by layering a novel fast graph representation learning layer between the CLIP and side networks. This integration introduces an inductive bias for the aggregation of local features, thereby better addressing the challenges in semantic segmentation. Through validation on five distinct datasets and extensive ablation studies, we have demonstrated the effectiveness of our modifications. Our findings indicate that with a slight increase in the number of parameters, there is a significant enhancement in performance.
In this paper, a robot visionrecognition system is developed based on the Robot Operating System (ROS) and the Open Source computervision (Open CV), which mainly implements face recognition, object detection, motion...
详细信息
ISBN:
(纸本)9781665464680
In this paper, a robot visionrecognition system is developed based on the Robot Operating System (ROS) and the Open Source computervision (Open CV), which mainly implements face recognition, object detection, motion analysis and object segmentation of the robot.
Time-series data mining plays an important role in big data decision making because it can reveal the development pattern of things. Similar concatenation of temporal data is a fundamental prerequisite for data twinni...
详细信息
Earth currently hosts 18 genera and 41 deer species, yet various human activities have degraded their natural habitats. According to the IUCN Red List, 10 deer species are now at a heightened risk of extinction. To de...
详细信息
ISBN:
(纸本)9798350386851;9798350386844
Earth currently hosts 18 genera and 41 deer species, yet various human activities have degraded their natural habitats. According to the IUCN Red List, 10 deer species are now at a heightened risk of extinction. To detect and conserve endangered deer populations, this work proposes a YOLOv9 model combined with the Squeeze-and-excitation networks (SENet) to build a Deer Species and Gender Detection System. Compared to other deer detection systems, the proposed system demonstrates the capability to do deer detection, deer species classification, and deer gender identification with minimal data, which has 94% mAP@0.5.
暂无评论