Table 2 Comparison with different models on Pascal VOC dataset.

From: LKC-Net: large kernel convolution object detection network

Models

mAP0.5 (%)

Params (M)

Fast YOLO18

52.7

–

Faster R-CNN-VGG-1618

73.2

–

SSD30018

79.6

36.1

SSD51218

81.6

36.1

SSD51218

81.6

36.1

ShuffleNetV2-SSDLite32

65.4

2.17

RefineDet512-VGG-1633

83.8

–

RFB Net512-VGG33

82.8

–

MobileNetV2-YOLOV434

81.5

46.34

EEEA-Net-C2-YOLOV434

81.8

31.15

YOLOV5s29

82.8*

7.06

YOLOV6-N35

79.3*

6.4

YOLOV7-Tiny36

80.9*

6.2

LKC-Net(Ours)

84.0

7.28

  1. The results with * in the table mean that this model is tested in the same environment, and the code comes from open source.
  2. Bold indicates the optimal performance.
Back to article page