Model with Spatially Weighted Pooling (SWP)

An imeplementation of Model+SWP using Keras (TF2) framework. Models supported: ResNet{50, 101}, VGG16, AlexNet

Implementation

ResNet_SWP

• base_model_trainable: bool = True
• input_shape:(224, 224, 3)
• swp_num_of_masks: 9
• fc_nodes: 1024
• weight_decay: 0.0005
• stddev: 0.005
• dropout_ratio: 0.5

You should set learning_rate of optimizers in {0.001, 0.0001, 0.00001}

VGG16_SWP

• base_model_trainable: bool = True
• input_shape:(224, 224, 3)
• swp_num_of_masks: 9
• fc_nodes: 512
• weight_decay: 0.0005
• stddev: 0.005
• dropout_ratio: 0.5

You should set learning_rate of optimizers in {0.001, 0.0001, 0.00001}

AlexNet_SWP

• base_model_trainable: bool = True
• input_shape:(227, 227, 3)
• swp_num_of_masks: 9
• fc_nodes: 512
• weight_decay: 0.0005
• stddev: 0.005
• dropout_ratio: 0.5

You should set learning_rate of optimizers in {0.001, 0.0001}

Pretrained model for ResNet50_SWP

I trained the CompCars dataset on ResNet50_SWP. To use it, use this command to combine files:

cat pretrained/resnet50_* > resnet50.h5

Citation

The idea is from two these papers (special thanks to the authors):

[https://ieeexplore.ieee.org/document/7891907]
Q. Hu, H. Wang, T. Li and C. Shen, "Deep CNNs With Spatially Weighted Pooling for Fine-Grained Car Recognition," in IEEE Transactions on Intelligent Transportation Systems, vol. 18, no. 11, pp. 3147-3156, Nov. 2017, doi: 10.1109/TITS.2017.2679114.

[https://arxiv.org/abs/1506.08959]
Q. Hu, H. Wang, T. Li and C. Shen, "Deep CNNs With Spatially Weighted Pooling for Fine-Grained Car Recognition," in IEEE Transactions on Intelligent Transportation Systems, vol. 18, no. 11, pp. 3147-3156, Nov. 2017, doi: 10.1109/TITS.2017.2679114.