{"id":21233538,"url":"https://github.com/midusi/handshape_datasets","last_synced_at":"2025-07-10T17:31:41.887Z","repository":{"id":53746114,"uuid":"161216100","full_name":"midusi/handshape_datasets","owner":"midusi","description":"A single library to (down)load all existing sign language handshape datasets.","archived":false,"fork":false,"pushed_at":"2024-12-18T17:12:44.000Z","size":6974,"stargazers_count":14,"open_issues_count":4,"forks_count":2,"subscribers_count":4,"default_branch":"master","last_synced_at":"2025-07-03T11:45:08.758Z","etag":null,"topics":["dataset","dataset-manager","download","handshape","handshape-datasets","python3","sign-language"],"latest_commit_sha":null,"homepage":"","language":"Python","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":"agpl-3.0","status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/midusi.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":"LICENSE","code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null}},"created_at":"2018-12-10T18:04:26.000Z","updated_at":"2024-12-18T17:12:47.000Z","dependencies_parsed_at":"2022-09-26T20:02:13.711Z","dependency_job_id":null,"html_url":"https://github.com/midusi/handshape_datasets","commit_stats":null,"previous_names":[],"tags_count":8,"template":false,"template_full_name":null,"purl":"pkg:github/midusi/handshape_datasets","repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/midusi%2Fhandshape_datasets","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/midusi%2Fhandshape_datasets/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/midusi%2Fhandshape_datasets/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/midusi%2Fhandshape_datasets/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/midusi","download_url":"https://codeload.github.com/midusi/handshape_datasets/tar.gz/refs/heads/master","sbom_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/midusi%2Fhandshape_datasets/sbom","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":264619050,"owners_count":23638397,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["dataset","dataset-manager","download","handshape","handshape-datasets","python3","sign-language"],"created_at":"2024-11-20T23:59:16.899Z","updated_at":"2025-07-10T17:31:36.878Z","avatar_url":"https://github.com/midusi.png","language":"Python","readme":"![handshape](http://facundoq.github.io/datasets/lsa16/files/dataset/1_1_1.png \"sample handshape\") \n![handshape](http://facundoq.github.io/datasets/lsa16/files/dataset/2_1_1.png \"sample handshape\")\n![handshape](http://facundoq.github.io/datasets/lsa16/files/dataset/3_1_1.png \"sample handshape\")\n\n# Goal\nThere are [various handshape datasets](http://facundoq.github.io/unlp/sign_language_datasets/) for Sign Language. However:\n* Each dataset has its own format and many are hard to find. \n* Each dataset has its own mapping of handshapes to classes. While signs depend on the specific Sign Language for a country/region, handshapes are universal. Hence, they could be shared between datasets/tasks. \n\nThis library aims to provide two main features:\n* A simplified API to download and load handshape datasets\n* A mapping between datasets so that datasets can be merged for training/testing models.\n\nThis library is a *work in progress*. Contributions are welcome. If you wish to add a dataset you can make a push request or open an issue.\n\n# Installation\n\nYou can install `handshape_datasets` via pip with:\n\n`pip install handshape_datasets`\n\n# Basic usage\n\nSimply import the module and load a dataset. The following downloads, preprocesses and load to memory the [LSA16 dataset](http://facundoq.github.io/datasets/lsa16/):\n\n    import handshape_datasets as hd\n    images,metadata = hd.load(\"lsa16\")\nAfterwards you can, for example, plot the first images of the dataset\n\n    import matplotlib.pyplot as plt\n    plt.imshow(images[0,:,:,:]) # show the first sample of the dataset\n\n# Advanced usage\n    import handshape_datasets as hd\n    hd.list_datasets() # List available datasets\n    hd.load(\"lsa16\",version=\"color\",delete=True) # use the color version, delete temporary files\n    hd.delete_temporary_files(\"lsa16\")# Delete temporary files  (if any)\n    hd.clear(\"lsa16\") # Delete all the local files for dataset LSA16\n    hd.info(\"lsa16\") # Shows detailed info of the dataset, including url, data format, fields, etc.\n    \n\n# Supported datasets\n\n\n|    Dataset id | Download size | Size on disk | Samples | Classes |\n|---------------|---------------|--------------|---------|---------|\n|     lsa16     |    640.6 Kb   |    1.2 Mb    |   800   |    16   |\n|      rwth     |    44.8 Mb    |   206.8 Mb   |   3359  |    45   |\n|     Irish     |    173.4 Mb   |   515.0 Mb   |  58114  |    26   |\n|     Ciarp     |    10.6 Mb    |   18.6 Mb    |   7127  |    10   |\n| PugeaultASL_A |     2.1 Gb    |    4.3 Gb    |  65774  |    24   |\n| PugeaultASL_B |    317.4 Mb   |   717.9 Mb   |  72676  |    26   |\n|    indianA    |     1.7 Gb    |    1.9 Gb    |   5040  |   140   |\n|    indianB    |    320.5 Mb   |    8.6 Gb    |   5000  |   140   |\n|      Nus1     |     2.8 Mb    |    3.6 Mb    |   479   |    10   |\n|      Nus2     |    73.7 Mb    |   106.1 Mb   |   2750  |    10   |\n|      jsl      |     4.5 Mb    |    7.9 Mb    |   8055  |    41   |\n|      psl      |    285.2 Mb   |    1.2 Gb    |   960   |    16   |\n\n\nYou can find more information about the datasets in the following [sign language dataset survey](http://facundoq.github.io/unlp/sign_language_datasets/)\n\n## Training a handshape classifier with Keras\n\nLoad the dataset:\n\n    x,metadata = handshape_datasets.load(\"lsa16\")\n    y = metadata[\"y\"]\nGet the input_shape and number of classes:\n\n    input_shape = x[0].shape\n    classes = y.max() + 1\n\nDefine a model (using a pretrained MobileNet here):\n\n    base_model = keras.applications.mobilenet.MobileNet(input_shape=(input_shape[0],input_shape[1],3), \n                                                                weights='imagenet', include_top=False)\n    output = keras.layers.GlobalAveragePooling2D()(base_model.output)\n    output = keras.layers.Dense(32, activation='relu')(output)\n    output = keras.layers.Dense(classes, activation='softmax')(output)\n    model = Model(inputs=base_model.input, outputs=output)\n    model.compile(optimizer='Adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])\n\nSplit the dataset intro train/test sets:\n\n    X_train, X_test, Y_train, Y_test = sklearn.model_selection.train_test_split(x,y,\n                                                                                    test_size=0.9,\n                                                                                    stratify=y)\n\nFit the model\n\n    history = model.fit(X_train, Y_train, batch_size=self.batch_size, epochs=self.epochs, validation_data=(X_test, Y_test))\n\n# Google Colab example:\n\nhttps://colab.research.google.com/drive/1kY-YrbegGFVT7NqVaeA4RjXYRVlZiISR?usp=sharing\n","funding_links":[],"categories":[],"sub_categories":[],"project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmidusi%2Fhandshape_datasets","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fmidusi%2Fhandshape_datasets","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fmidusi%2Fhandshape_datasets/lists"}