iMTE

논문 제목 : Group-CAM: Group Score-Weighted Visual Explanations for Deep Convolutional Networks 논문 주소 : https://arxiv.org/abs/2103.13859 Group-CAM: Group Score-Weighted Visual Explanations for Deep Convolutional Networks In this paper, we propose an efficient saliency map generation method, called Group score-weighted Class Activation Mapping (Group-CAM), which adopts the "split-transform-merge" str..

논문 제목 : SS-CAM: Smoothed Score-CAM for Sharper Visual Feature Localization 논문 주소 : https://arxiv.org/abs/2006.14255 SS-CAM: Smoothed Score-CAM for Sharper Visual Feature Localization Interpretation of the underlying mechanisms of Deep Convolutional Neural Networks has become an important aspect of research in the field of deep learning due to their applications in high-risk environments. To expl..

논문 제목 : Score-CAM : Score-weighted visual explanations for convolutional neural networks 논문 주소 : https://openaccess.thecvf.com/content_CVPRW_2020/html/w1/Wang_Score-CAM_Score-Weighted_Visual_Explanations_for_Convolutional_Neural_Networks_CVPRW_2020_paper.html CVPR 2020 Open Access Repository Haofan Wang, Zifan Wang, Mengnan Du, Fan Yang, Zijian Zhang, Sirui Ding, Piotr Mardziel, Xia Hu; Proceedi..

논문 제목 : Adapting Grad-CAM for Embedding Networks 논문 주소 : https://openaccess.thecvf.com/content_WACV_2020/html/Chen_Adapting_Grad-CAM_for_Embedding_Networks_WACV_2020_paper.html WACV 2020 Open Access Repository Lei Chen, Jianhui Chen, Hossein Hajimirsadeghi, Greg Mori; Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2020, pp. 2794-2803 The gradient-weighte..

논문 제목 : Interpretable and fine-grained visual explanations for CNNs 논문 주소 : openaccess.thecvf.com/content_CVPR_2019/html/Wagner_Interpretable_and_Fine-Grained_Visual_Explanations_for_Convolutional_Neural_Networks_CVPR_2019_paper.html CVPR 2019 Open Access Repository Interpretable and Fine-Grained Visual Explanations for Convolutional Neural Networks Jorg Wagner, Jan Mathias Kohler, Tobias Gindel..

논문 제목 : Sanity checks for saliency maps 논문 주소 : arxiv.org/abs/1810.03292 Sanity Checks for Saliency Maps Saliency methods have emerged as a popular tool to highlight features in an input deemed relevant for the prediction of a learned model. Several saliency methods have been proposed, often guided by visual appeal on image data. In this work, we propose an a arxiv.org 주요 내용 : 1) Saliency map은 학..

논문 제목 : Sanity checks for saliency maps 논문 주소 : arxiv.org/abs/1810.03292 Sanity Checks for Saliency Maps Saliency methods have emerged as a popular tool to highlight features in an input deemed relevant for the prediction of a learned model. Several saliency methods have been proposed, often guided by visual appeal on image data. In this work, we propose an a arxiv.org 주요 수식 정리: 0) Definition in..

논문 제목 : SmoothGrad : removing noise by adding noise 논문 주소 : arxiv.org/abs/1706.03825 SmoothGrad: removing noise by adding noise Explaining the output of a deep network remains a challenge. In the case of an image classifier, one type of explanation is to identify pixels that strongly influence the final decision. A starting point for this strategy is the gradient of the class score arxiv.org 주요 ..

논문 제목 : Smooth Grad-CAM++: An Enhanced Inference Level Visualization Technique for Deep Convolutional Neural Network Models 논문 주소 : arxiv.org/abs/1908.01224 Smooth Grad-CAM++: An Enhanced Inference Level Visualization Technique for Deep Convolutional Neural Network Models Gaining insight into how deep convolutional neural network models perform image classification and how to explain their outpu..

논문 제목 : Grad-CAM++: Generalized Gradient-based Visual Explanations for Deep Convolutional Networks 논문 주소 : arxiv.org/pdf/1710.11063.pdf IEEE WACV (2018, ieeexplore.ieee.org/document/8354201)에 나온 논문을 바탕으로 이해하고 내용을 작성한다. arixv에서 나온 버전이 좀 더 extended version임으로 Grad-CAM++에 더 깊은 이해를 위해서는 extended version을 읽는 것을 추천한다. 주요 내용 : 1) Deep models은 "black box"로서 internal function을 이해하는데에는 어려움이 있다. 이를 해결하기 위해..