Description
ResNet-D is a modification on the ResNet architecture that utilises an average pooling tweak for downsampling. The motivation is that in the unmodified ResNet, the 1 × 1 convolution for the downsampling block ignores 3/4 of input feature maps, so this is modified so no information will be ignored
Papers Using This Method
Optimizing Anchor-based Detectors for Autonomous Driving Scenes2022-08-11Revisiting 3D ResNets for Video Recognition2021-09-03Simple Training Strategies and Model Scaling for Object Detection2021-06-30Revisiting ResNets: Improved Training and Scaling Strategies2021-03-13Prediction of Chronic Kidney Disease Using Deep Neural Network2020-12-22Real-time object detection method based on improved YOLOv4-tiny2020-11-09WaveGrad: Estimating Gradients for Waveform Generation2020-09-02Compounding the Performance Improvements of Assembled Techniques in a Convolutional Neural Network2020-01-17W-Net: A CNN-based Architecture for White Blood Cells Image Classification2019-10-02Bag of Tricks for Image Classification with Convolutional Neural Networks2018-12-04