TR-ICANet: A Fast Unsupervised Deep-Learning-Based Scheme for Unconstrained Ear Recognition

Arabian Journal for Science and Engineering, Journal Year: 2022, Volume and Issue: 47(8), P. 9887 - 9898

Published: Jan. 9, 2022

Language: Английский

---

A Systematic Literature Review on Human Ear Biometrics: Approaches, Algorithms, and Trend in the Last Decade

Information, Journal Year: 2023, Volume and Issue: 14(3), P. 192 - 192

Published: March 17, 2023

Biometric technology is fast gaining pace as a veritable developmental tool. So far, biometric procedures have been predominantly used to ensure identity and ear recognition techniques continue provide very robust research prospects. This paper proposes identify review present for biometrics using certain parameters: machine learning methods, directions future research. Ten databases were accessed, including ACM, Wiley, IEEE, Springer, Emerald, Elsevier, Sage, MIT, Taylor & Francis, Science Direct, 1121 publications retrieved. In order obtain relevant materials, some articles excused criteria such abstract eligibility, duplicity, uncertainty (indeterminate method). As result, 73 papers selected in-depth assessment significance. A quantitative analysis was carried out on the identified works search strategies: source, technique, datasets, status, architecture. Quantitative Analysis (QA) of feature extraction methods studies with geometric approach indicating highest value at 36%, followed by local method 27%. Several architectures, Convolutional Neural Network, restricted Boltzmann machine, auto-encoder, deep belief network, other unspecified showed 38%, 28%, 21%, 5%, 4%, respectively. Essentially, this survey also provides various status existing in classifying related studies. taxonomy current methodologies system presented along publicly available occlussion pose sensitive black image dataset 970 images. The study concludes need researchers consider improvements speed security algorithms.

Language: Английский

---

Synthesizing multilevel abstraction ear sketches for enhanced biometric recognition

Image and Vision Computing, Journal Year: 2025, Volume and Issue: unknown, P. 105424 - 105424

Published: Jan. 1, 2025

Language: Английский

---

Research on herd sheep facial recognition based on multi-dimensional feature information fusion technology in complex environment

Frontiers in Veterinary Science, Journal Year: 2025, Volume and Issue: 12

Published: March 13, 2025

Intelligent management of large-scale farms necessitates efficient monitoring individual livestock. To address this need, a three-phase intelligent system based on deep learning was designed, integrating multi-part detection network for flock inventory counting, facial classification model identity recognition, and expression analysis health assessment. For network, The YOLOv5s path aggregation modified by incorporating multi-link convolution fusion block (MCFB) to enhance fine-grained feature extraction across objects different sizes. improve the dense small targets, Re-Parameterizable Convolution (RepConv) structure introduced into head. sixth-stage in GhostNet replaced with four-layer spatially separable self-attention mechanism (SSSA) strengthen key extraction. Additionally, compression techniques were applied optimize improved efficiency. A transfer strategy employed weight pre-training, performance evaluated using FPS, weight, mean average precision (mAP), test set accuracy. Experimental results demonstrated that enhanced identification effectively extracted features from regions sheep flock, achieving an accuracy 95.84%, 2.55% improvement mAP compared YOLOv5s. achieved 98.9%, surpassing 3.1%. attained 99.2%, representing 3.6% increase EfficientNet. proposed significantly enhances efficiency advanced optimization techniques. improvements further enable real-time monitoring, contributing livestock management.

Language: Английский

---

Enhanced Plant Health Monitoring with Dual Head CNN for Leaf Classification and Disease Identification

Journal of Agriculture and Food Research, Journal Year: 2025, Volume and Issue: unknown, P. 101930 - 101930

Published: April 1, 2025

Language: Английский

---