最近の記事

ffmpegによる動画のクロッピング

ffmpeg -i 変換前ファイル -vf crop= 出力動画の幅 : 出力動画の高さ : 元動画の左上をゼロとして X 軸の距離 : 元動画の左上をゼロとして Y 軸の距離 動画の左上が原点 ffmpeg -i before.mpg -vf crop=200:150:100:50 after.mpg

    • Pytorch CustomDatasetの作り方

      torch.utils.data.Datasetを継承 class CustomDataset(torch.utils.data.Dataset): def __init__(self): #読み込むファイルパス一覧とラベル一覧を用意 self.images = self.labels = def __getitem__(self, index): #画像をがいるパスから取り出す

      • Deep Metric Learning via Lifted Structured Feature Embedding

        Oh Song, Hyun, et al. "Deep metric learning via lifted structured feature embedding." Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 2016. 目的 既存の手法 ・Embedding Spaceにおいて似ている画像は近くに,似ていない画像は遠くに配置 既存の手法の問題点(多クラス

        • BBeep: A Sonic Collision Avoidance System for Blind Travellers and Nearby Pedestrians

          Seita Kayukawa, Keita Higuchi, João Guerreiro, Shigeo Morishima, Yoichi Sato, Kris Kitani, and Chieko Asakawa. 2019. BBeep: A Sonic Collision Avoidance System for Blind Travellers and Nearby Pedestrians. In Proceedings of the 2019 CHI Confe

        ffmpegによる動画のクロッピング

          HindSight: Enhancing Spatial Awareness by Sonifying Detected Objects in Real-Time 360-Degree Video

          Eldon Schoop, James Smith, and Bjoern Hartmann. 2018. HindSight: Enhancing Spatial Awareness by Sonifying Detected Objects in Real-Time 360-Degree Video. In Proceedings of the 2018 CHI Conference on Human Factors in Computing Systems (CHI '

          HindSight: Enhancing Spatial Awareness by Sonifying Detected Objects in Real-Time 360-Degree Video

          Omnidirectional CNN for Visual Place Recognition and Navigation

          T. Wang, H. Huang, J. Lin, C. Hu, K. Zeng and M. Sun, "Omnidirectional CNN for Visual Place Recognition and Navigation," 2018 IEEE International Conference on Robotics and Automation (ICRA), Brisbane, QLD, 2018, pp. 2341-2348. doi: 10.1109/

          Omnidirectional CNN for Visual Place Recognition and Navigation

          Indoor Scene Layout Estimation from a Single Image

          H. J. Lin, S. Huang, S. Lai and C. Chiang, "Indoor Scene Layout Estimation from a Single Image," 2018 24th International Conference on Pattern Recognition (ICPR), Beijing, 2018, pp. 842-847. doi: 10.1109/ICPR.2018.8546278 keywords: {augment

          Indoor Scene Layout Estimation from a Single Image

          Lifelogging Retrieval based on Semantic Concepts Fusion

          Thanh-Dat Truong, Tung Dinh-Duy, Vinh-Tiep Nguyen, and Minh-Triet Tran. 2018. Lifelogging Retrieval based on Semantic Concepts Fusion. In Proceedings of the 2018 ACM Workshop on The Lifelog Search Challenge (LSC '18). ACM, New York, NY, USA

          Lifelogging Retrieval based on Semantic Concepts Fusion

          Unsupervised Visual Representation Learning for Indoor Scenes with a Siamese ConvNet and Graph Constraints

          Liu, M.; Chen, R.; Ai, H.; Chen, Y.; Li, D. Unsupervised Visual Representation Learning for Indoor Scenes with a Siamese ConvNet and Graph Constraints. Preprints 2019, 2019030189 目的 ・Indoor Sceneは多くのカテゴリやカテゴリ間の類似性によって,視覚的表現を効果的に学習するのが難しい ・

          Unsupervised Visual Representation Learning for Indoor Scenes with a Siamese ConvNet and Graph Constraints

          Approximating CNNs with Bag-of-local-Features models works surprisingly well on ImageNet

          Brendel, Wieland, and Matthias Bethge. "Approximating cnns with bag-of-local-features models works surprisingly well on imagenet." arXiv preprint arXiv:1904.00760 (2019). 目的 空間的な順序を考慮せずに,画像の局所的特徴に基づいて画像を分類したい 具体的手法 ・Bag of Feature ・Bag of

          Approximating CNNs with Bag-of-local-Features models works surprisingly well on ImageNet

          Bag of Visual Words

          目的 画像データセットの局所特徴量を基に,各画像の特徴ベクトルを作成する手法 具体的手法 全画像から局所特徴ベクトルを抽出 (CNNならCNNの中間層から抽出) それらの局所特徴ベクトルをクラスタリング(k-NNなど) 得られたクラスタ中心をVisual Wordsとする 各画像についてVisual Wordsとなった局所特徴ベクトルの出現頻度について,ヒストグラム化する->これがその画像の特徴ベクトルとなる =Bag of Visual Words 実装 Bag of

          Bag of Visual Words

          Github Error remote: Invalid username or password

          remote set-url でリポジトリを上書きする git remote set-url origin {SSH URLs}

          Github Error remote: Invalid username or password

          scipy.sparse の扱いかた

          l = [[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 12], [13, 14, 15, 16]]csr = scipy.sparse.csr_matrix(l)print(csr[1])# (0, 0) 5# (0, 1) 6# (0, 2) 7# (0, 3) 8 デフォルトでは行の要素が出てくる l = [[1, 2, 3, 4], [5, 6, 7, 8], [9, 10, 11, 1

          scipy.sparse の扱いかた