{
  "fetched_at": "2026-10-19T21:37:55Z",
  "source_url": "https://www.kaggle.com/competitions/dsg-challenge-1-supervised-learning/data",
  "pages": [
    {
      "name": "data-description",
      "content": "## Dataset Details:\nThe dataset includes:\n\n- Seven videos of people making different hand gestures.\n- For each video, two files containing 50 carefully chosen attributes that describe the gestures\n\n# Gesture Phase Segmentation Dataset\n\n## Attribute Information:\n\n### Raw Files:\n\n1. `lhx` - Position of left hand (x coordinate)\n2. `lhy` - Position of left hand (y coordinate)\n3. `lhz` - Position of left hand (z coordinate)\n4. `rhx` - Position of right hand (x coordinate)\n5. `rhy` - Position of right hand (y coordinate)\n6. `rhz` - Position of right hand (z coordinate)\n7. `hx` - Position of head (x coordinate)\n8. `hy` - Position of head (y coordinate)\n9. `hz` - Position of head (z coordinate)\n10. `sx` - Position of spine (x coordinate)\n11. `sy` - Position of spine (y coordinate)\n12. `sz` - Position of spine (z coordinate)\n13. `lwx` - Position of left wrist (x coordinate)\n14. `lwy` - Position of left wrist (y coordinate)\n15. `lwz` - Position of left wrist (z coordinate)\n16. `rwx` - Position of right wrist (x coordinate)\n17. `rwy` - Position of right wrist (y coordinate)\n18. `rwz` - Position of right wrist (z coordinate)\n19. `timestamp` - Timestamp of the frame\n20. `phase`: \n    - Rest\n    - Preparation\n    - Stroke\n    - Hold\n    - Retraction\n\n### Processed Files:\n\n1. Vectorial velocity of left hand (x coordinate)\n2. Vectorial velocity of left hand (y coordinate)\n3. Vectorial velocity of left hand (z coordinate)\n4. Vectorial velocity of right hand (x coordinate)\n5. Vectorial velocity of right hand (y coordinate)\n6. Vectorial velocity of right hand (z coordinate)\n7. Vectorial velocity of left wrist (x coordinate)\n8. Vectorial velocity of left wrist (y coordinate)\n9. Vectorial velocity of left wrist (z coordinate)\n10. Vectorial velocity of right wrist (x coordinate)\n11. Vectorial velocity of right wrist (y coordinate)\n12. Vectorial velocity of right wrist (z coordinate)\n13. Vectorial acceleration of left hand (x coordinate)\n14. Vectorial acceleration of left hand (y coordinate)\n15. Vectorial acceleration of left hand (z coordinate)\n16. Vectorial acceleration of right hand (x coordinate)\n17. Vectorial acceleration of right hand (y coordinate)\n18. Vectorial acceleration of right hand (z coordinate)\n19. Vectorial acceleration of left wrist (x coordinate)\n20. Vectorial acceleration of left wrist (y coordinate)\n21. Vectorial acceleration of left wrist (z coordinate)\n22. Vectorial acceleration of right wrist (x coordinate)\n23. Vectorial acceleration of right wrist (y coordinate)\n24. Vectorial acceleration of right wrist (z coordinate)\n25. Scalar velocity of left hand\n26. Scalar velocity of right hand\n27. Scalar velocity of left wrist\n28. Scalar velocity of right wrist\n29. Scalar acceleration of left hand\n30. Scalar acceleration of right hand\n31. Scalar acceleration of left wrist\n32. Scalar acceleration of right wrist\n33. `phase`:\n    - D (rest position, from Portuguese \"descanso\")\n    - P (preparation)\n    - S (stroke)\n    - H (hold)\n    - R (retraction)\n\n## Acknowledgement\n\nThe Gesture Phase Segmentation Dataset consists of temporal segmentation of gestures performed by gesture researchers to preprocess videos for further analysis. The description of the gesture phase segmentation problem may be found in MADEO et al. (2013a) and MADEO et al. (2013b).\n\nThe dataset includes seven videos recorded using Microsoft Kinect sensors. Three different users were asked to read three comic strips and tell the stories in front of the sensor. For each frame, the dataset contains an image identified by a timestamp and a text file containing positions (coordinates x, y, z) of six articulation points - left hand, right hand, left wrist, right wrist, head, and spine - with each line corresponding to a frame and identified by a timestamp. Manual segmentation of each file by a specialist provided a ground truth for classification.\n\nThe dataset is organized into 14 files, seven raw files, and seven processed files, one for each video in the dataset. The name of each file refers to the corresponding video, with the letter corresponding to the user (A, B, C) and the number corresponding to the story (1, 2, 3). The raw files contain information obtained from Microsoft Kinect, as described above. The processed file contains vectorial and scalar velocity and acceleration of left hand, right hand, left wrist, and right wrist. These measures were obtained after normalizing positions of the hands and wrists relative to the position of the head and spine, using a displacement of 3 to measure velocity, as described in MADEO et al. (2013c).\n\"\"\"\n## Citation\n>Madeo,Renata, Wagner,Priscilla, and Peres,Sarajane. (2014). Gesture Phase Segmentation. UCI Machine Learning Repository. https://doi.org/10.24432/C5Z32C.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Description",
      "content": "## Description\n\nWelcome to the \"Supervised Learning Challenge\", the First of our many upcoming challenges! In this competition, you'll use a unique dataset of features extracted from seven videos. Your task is to build machine learning models to identify different gesture phases in these videos.\n\nYou can get the data in the data section.\n### ANY ADDITIONAL FINDINGS FROM THE DATA WILL BE DULY REWARDED",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "Evaluation",
      "content": "## Tasks:\n- Gesture Classification: Label each segmented part with the type of gesture it represents\n- This is an open-ended challenge \n\n## Bonus Task\n- Cluster the data to find various hand movements/positions through unsupervised learning.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "rules",
      "content": "HAVE FUN!",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    },
    {
      "name": "SUBMISSION GUIDELINES",
      "content": "## GUIDELINES\n1. The solution is customary ```test.csv``` data is NOT the raw data, it is processed data.\n2. You are expected to apply multiple approaches and perform Exploratory Analysis.\n3. Compile all the results in the beginning of the notebook in a markdown file.",
      "isPublished": false,
      "hasIsPublished": false,
      "postTitle": "",
      "hasPostTitle": false,
      "mimeType": "",
      "hasMimeType": false
    }
  ]
}