Test results of different models on TinyPerson. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”

Differences between models of different scales. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

Visual comparison of TinyPerson. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”

LC-FPN structure. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

Labeling information of the VisDrone dataset. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

LFERELAN structure. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

The experimental environment. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

LCFF-Net network structure. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

LDSCD-Head structure. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

Ablation experiment result on VisDrone-val. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

The key parameter configurations. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

LR-NET structure. by Daoze Tang (20454615)

“…To address these issues, we propose an improved, lightweight algorithm: LCFF-Net. First, we propose the LFERELAN module, designed to enhance the extraction of tiny target features and optimize the use of computational resources. …”

Test results of different models on VisDrone2019. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”

Structure of YOLOv11 network. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”

Structure of SEAM network [16]. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”

Ablation experiment curve. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”

Experimental environment configuration. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”

Architecture of BiFPN network [14]. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”

Datasets label distribution map. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”

Structure of UAS-YOLO network. by Ju Liang (4277062)

“…These characteristics impose high demands on detection algorithms in terms of fine-grained feature extraction, cross-scale fusion capability, and occlusion resistance.The YOLOv11s model has significant limitations in practical applications: its feature extraction module has a single semantic representation, the traditional feature pyramid network has limited capability to detect multi-scale targets, and it lacks an effective feature compensation mechanism when targets are occluded.To address these issues, we propose a UAV aerial small target detection algorithm named UAS-YOLO (Universal Inverted Bottleneck with Adaptive BiFPN and Separated and Enhancement Attention module YOLO), which incorporates three key optimizations. …”