阴道镜序列（视频）中随机抽取图像数据集

Kelwin Fernandes (kafc _at_ inesctec _dot_ pt) - INESC TEC & FEUP, Porto, Portugal. Jaime S. Cardoso - INESC TEC & FEUP, Porto, Portugal. Jessica Fernandes - Universidad Central de Venezuela, Caracas, Venezuela. Data Set Information: * The dataset was acquired and annotated by professional physicians at 'Hospital Universitario de Caracas'. * The subjective judgments (target variables) were originally done in an ordinal manner (poor, fair, good, excellent) and was discretized in two classes (bad, good). * Images were randomly sampled from the original colposcopic sequences (videos). * The original images and the manual segmentations are included in the 'images' directory. * The dataset has three modalities (i.e. Hinselmann, Green, Schiller). * The target variables are expert::X (X in 0,...,5) and consensus. Attribute Information: Three modalities: hinselmann, green, schiller. Number of Attributes: 69 (62 predictive attributes, 7 target variables) cervix_area: image area with cervix. os_area: image area with external os. walls_area: image area with vaginal walls. speculum_area: image area with the speculum. artifacts_area: image area with artifacts. cervix_artifacts_area: cervix area with the artifacts. os_artifacts_area: external os area with the artifacts. walls_artifacts_area: vaginal walls with the artifacts. speculum_artifacts_area: speculum area with the artifacts. cervix_specularities_area: cervix area with the specular reflections. os_specularities_area: external os area with the specular reflections. walls_specularities_area: vaginal walls area with the specular reflections. speculum_specularities_area: speculum area with the specular reflections. specularities_area: total area with specular reflections. area_h_max_diff: maximum area differences between the four cervix quadrants. rgb_cervix_r_mean: average color information in the cervix (R channel). rgb_cervix_r_std: stddev color information in the cervix (R channel). rgb_cervix_r_mean_minus_std: (avg - stddev) color information in the cervix (R channel). rgb_cervix_r_mean_plus_std: (avg + stddev) information in the cervix (R channel). rgb_cervix_g_mean: average color information in the cervix (G channel). rgb_cervix_g_std: stddev color information in the cervix (G channel). rgb_cervix_g_mean_minus_std: (avg - stddev) color information in the cervix (G channel). rgb_cervix_g_mean_plus_std: (avg + stddev) color information in the cervix (G channel). rgb_cervix_b_mean: average color information in the cervix (B channel). rgb_cervix_b_std: stddev color information in the cervix (B channel). rgb_cervix_b_mean_minus_std: (avg - stddev) color information in the cervix (B channel). rgb_cervix_b_mean_plus_std: (avg + stddev) color information in the cervix (B channel). rgb_total_r_mean: average color information in the image (B channel). rgb_total_r_std: stddev color information in the image (R channel). rgb_total_r_mean_minus_std: (avg - stddev) color information in the image (R channel). rgb_total_r_mean_plus_std: (avg + stddev) color information in the image (R channel). rgb_total_g_mean: average color information in the image (G channel). rgb_total_g_std: stddev color information in the image (G channel). rgb_total_g_mean_minus_std: (avg - stddev) color information in the image (G channel). rgb_total_g_mean_plus_std: (avg + stddev) color information in the image (G channel). rgb_total_b_mean: average color information in the image (B channel). rgb_total_b_std: stddev color information in the image (B channel). rgb_total_b_mean_minus_std: (avg - stddev) color information in the image (B channel). rgb_total_b_mean_plus_std: (avg + stddev) color information in the image (B channel). hsv_cervix_h_mean: average color information in the cervix (H channel). hsv_cervix_h_std: stddev color information in the cervix (H channel). hsv_cervix_s_mean: average color information in the cervix (S channel). hsv_cervix_s_std: stddev color information in the cervix (S channel). hsv_cervix_v_mean: average color information in the cervix (V channel). hsv_cervix_v_std: stddev color information in the cervix (V channel). hsv_total_h_mean: average color information in the image (H channel). hsv_total_h_std: stddev color information in the image (H channel). hsv_total_s_mean: average color information in the image (S channel). hsv_total_s_std: stddev color information in the image (S channel). hsv_total_v_mean: average color information in the image (V channel). hsv_total_v_std: stddev color information in the image (V channel). fit_cervix_hull_rate: Coverage of the cervix convex hull by the cervix. fit_cervix_hull_total: Image coverage of the cervix convex hull. fit_cervix_bbox_rate: Coverage of the cervix bounding box by the cervix. fit_cervix_bbox_total: Image coverage of the cervix bounding box. fit_circle_rate: Coverage of the cervix circle by the cervix. fit_circle_total: Image coverage of the cervix circle. fit_ellipse_rate: Coverage of the cervix ellipse by the cervix. fit_ellipse_total: Image coverage of the cervix ellipse. fit_ellipse_goodness: Goodness of the ellipse fitting. dist_to_center_cervix: Distance between the cervix center and the image center. dist_to_center_os: Distance between the cervical os center and the image center. experts::0: subjective assessment of the Expert 0 (target variable). experts::1: subjective assessment of the Expert 1 (target variable). experts::2: subjective assessment of the Expert 2 (target variable). experts::3: subjective assessment of the Expert 3 (target variable). experts::4: subjective assessment of the Expert 4 (target variable). experts::5: subjective assessment of the Expert 5 (target variable). consensus: subjective assessment of the consensus (target variable). Relevant Papers: Fernandes, Kelwin, Jaime S. Cardoso, and Jessica Fernandes. 'Transfer Learning with Partial Observability Applied to Cervical Cancer Screening.' Iberian Conference on Pattern Recognition and Image Analysis. Springer International Publishing, 2017. Citation Request: Fernandes, Kelwin, Jaime S. Cardoso, and Jessica Fernandes. 'Transfer Learning with Partial Observability Applied to Cervical Cancer Screening.' Iberian Conference on Pattern Recognition and Image Analysis. Springer International Publishing, 2017.

搜集汇总

背景与挑战

背景概述

该数据集包含从阴道镜视频中随机抽取的图像，具有三种模态和69个属性，由专业医师标注，适用于宫颈癌筛查研究。