Open chullhwan-song opened 5 years ago
Google Landmark Classification Challenge 와 헷갈리지 마세요~
구조
간단히 표현한다면
1th 에 사용한 REMAP Feature 에 대한 정체 힌트
https://twitter.com/ducha_aiki/status/1008833406036107270
multi-conv layer(계층별)에서 roi 별 aggregate한 feature를 만드는듯
Practically, the KL-divergence Weighting (KLW) block in the REMAP architecture is implemented
using a convolutional layer with weights initialized by the KL-divergence values and optimized
using Stochastic Gradient Descent (SGD) on the triplet loss function.
...
Our novel component KL-divergence weighting (KLW) can be implemented using 1D convolutional
layer, with weights than can be optimized.
1. Extract features from hidden layer
2. For each image from test set find K closest images from train set (K=5)
3. For each class_id we computed: scores[class_id] = sum(cos(query_image, index) for index in K_closest_images)
4. For each class_id we normalized its score: scores[class_id] /= min(K, number of samples in train dataset with class=class_id)
5. label = argmax(scores), confidence = scores[label]
1. Compute confidence score for each label using predictions from steps 1-4 as follow: score[label] = label_count / models_count + sum(label_confidence for each model) / models_count. Here label_count is a number of models where the prediction with max confidence is equal to the label.
2. We also used each prediction from step 5 with confidence = 1 + confidence_from_step_5 / 100
2 stage
소스 : https://github.com/jandaldrop/landmark-recognition-challenge/
https://www.kaggle.com/c/landmark-retrieval-challenge