Deep Learning-Based Object Detection Algorithms

One of the main areas of study in computer vision is object detection. It can identify the type and location of target items and determine whether they are present in pictures or movies. With the development of deep learning, Object detection algorithms have seen significant enhancements in both spe...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:ITM web of conferences Ročník 73; s. 02024
Hlavný autor: Yao, Linxi
Médium: Konferenčný príspevok.. Journal Article
Jazyk:English
Vydavateľské údaje: Les Ulis EDP Sciences 01.01.2025
Predmet:
Algorithms
Computer vision
Deep learning
Machine learning
Object recognition
Performance evaluation
Target detection
ISSN:2431-7578, 2271-2097
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:One of the main areas of study in computer vision is object detection. It can identify the type and location of target items and determine whether they are present in pictures or movies. With the development of deep learning, Object detection algorithms have seen significant enhancements in both speed and accuracy, leading to extensive adoption across various domains, including autonomous driving, drone surveillance, and security monitoring. This article examines some of the most well-known algorithms from the deep learning period, classifies them into four types of object identification algorithms—two-stage, one-stage, keypoint-based, and transformer-based — and describes their primary advances, benefits, and drawbacks. Furthermore, this work organizes target detection datasets and performance evaluation indicators that are routinely used in studies and provides detailed explanations of their content and properties. The paper adds to the study and advancement of target detection technology-related domains and serves as a useful resource for practitioners and scholars.
Bibliografia:ObjectType-Conference Proceeding-1
SourceType-Conference Papers & Proceedings-1
content type line 21
ISSN:2431-7578
2271-2097
DOI:10.1051/itmconf/20257302024