About this document
Efficient CNN for Aerial Crowd Counting by shrines.sj is a document available to read on EtoBox.
The document presents Flounder-Net, an efficient CNN model designed for crowd counting using aerial photography, which addresses challenges related to high-resolution images and limited computing resources. The model employs interleaved group convolution to reduce redundancy and enhance processing speed, achieving FCN-level accuracy while being 20% faster and having 17% fewer parameters than traditional models. Extensive experiments demonstrate its effectiveness across various camera types, including handhe
- Author
- shrines.sj
- Language
- EN