Object Detection Based on VGG with ResNet Network
Md Foysal Haque, Hye-Youn Lim, Dae-Seong Kang · 2019 International Conference on Electronics, Information, and Communication (ICEIC) · 2019
We introduce an improved very deep convolutional network for accurate and significant object detection. It extracts high-level features that help to achieve tremendous performance to classify the image and detecting objects. The very deep convolutional neural network classifies data very profoundly. The VGG network constructed based on the very deep convolutional neural network. To constructing VGG network a stack of small convolutional filters used. The VGG network achieved high accuracy to large-scale image classification, but it has some training and localization problem. To overcome this, we proposed an improved method that able to solve the problem. We implement ResNet network in VGG that minimize VGG networks training error and able to detect more small objects.