Data

Tools

Indexes

Applications

Experiments

Open Source Science

Recent Releases of zensvi

zensvi - v1.3.0

What's Changed

  • Enhance object detection with flexible output options and grouped summaries by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/119
  • Enhance image transformation with alpha channel support for sky view factor calculation by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/122
  • Update dependencies and package versions in poetry.lock and pyproject.toml by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/124
  • Enhacement/simplifydevicesetting by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/126
  • Add DepthAnythingV2 module with models and assets by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/129

Full Changelog: https://github.com/koito19960406/ZenSVI/compare/v1.2.0...v1.3.0

- Python
Published by koito19960406 9 months ago

zensvi - v1.2.0

What's Changed

  • Added metadata update capability for GSV by @koito19960406 in 366b9afe3bab43f0a16f1f5d62ef56a6c04c9734
    • New update_metadata method for GSVDownloader
    • Allows updating year/month metadata for existing panorama IDs without redoing the entire download process
    • Includes comprehensive test coverage

Contributors

  • @koito19960406

Full Changelog: https://github.com/koito19960406/ZenSVI/compare/v1.1.0...v1.2.0

- Python
Published by koito19960406 about 1 year ago

zensvi - v1.1.0

What's Changed

  • Update README.md by @seshing in https://github.com/koito19960406/ZenSVI/pull/105
  • Update poetry.lock with new dependencies and version upgrades by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/108
  • Remove deprecated modules and update dependency management by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/107
  • fix kartaview download api, add retry decorator by @Junguin in https://github.com/koito19960406/ZenSVI/pull/111
  • Use Path() for path_pid existence check by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/113
  • V1.0.2 by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/109
  • Enhance pixel ratio calculation in segmentation with improved normalization by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/115
  • Enhancement/object_detection by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/114
  • Added Jupyter notebooks of download, metadata, transform and visualiz… by @ZhuYihan-UMI in https://github.com/koito19960406/ZenSVI/pull/117
  • Enhancement/verbosity by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/118
  • V1.1.0 by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/116

New Contributors

  • @seshing made their first contribution in https://github.com/koito19960406/ZenSVI/pull/105
  • @ZhuYihan-UMI made their first contribution in https://github.com/koito19960406/ZenSVI/pull/117

Full Changelog: https://github.com/koito19960406/ZenSVI/compare/v.1.0.0...v1.1.0

- Python
Published by koito19960406 about 1 year ago

zensvi - v.1.0.0

What's Changed

  • Bugfix/mapillary kwargs by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/69
  • Svi embeddings by @MahmoudAbdelRahman in https://github.com/koito19960406/ZenSVI/pull/65
  • update-AMSDownloader-add-test_ams.py by @ruirzma in https://github.com/koito19960406/ZenSVI/pull/70
  • added point cloud by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/72
  • add kartaview download class by @Junguin in https://github.com/koito19960406/ZenSVI/pull/67
  • Enhancement/automate log by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/73
  • added a new argument for using only upper half by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/74
  • V1.0.0 point cloud modification by @fzc961020 in https://github.com/koito19960406/ZenSVI/pull/75
  • removed geopolars dependency by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/76
  • New features by @matqr in https://github.com/koito19960406/ZenSVI/pull/82
  • Add global streetscapes demo tutorial by @Junguin in https://github.com/koito19960406/ZenSVI/pull/90
  • Additional perception model and update test script by @matqr in https://github.com/koito19960406/ZenSVI/pull/91
  • Adding Exampe Notebook for Depth Estimation and Point Cloud Reconstru… by @fzc961020 in https://github.com/koito19960406/ZenSVI/pull/93
  • added embeddings tutorial by @MahmoudAbdelRahman in https://github.com/koito19960406/ZenSVI/pull/95
  • adding perception demo for xiuchengs and jianis model by @matqr in https://github.com/koito19960406/ZenSVI/pull/97
  • Integrate Jianis Perception Model and Enhance Testing Framework by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/101
  • Enhance documentation and improve code quality by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/103
  • Update documentation and improve code quality by @koito19960406 in https://github.com/koito19960406/ZenSVI/pull/104

New Contributors

  • @MahmoudAbdelRahman made their first contribution in https://github.com/koito19960406/ZenSVI/pull/65
  • @ruirzma made their first contribution in https://github.com/koito19960406/ZenSVI/pull/70
  • @Junguin made their first contribution in https://github.com/koito19960406/ZenSVI/pull/67
  • @fzc961020 made their first contribution in https://github.com/koito19960406/ZenSVI/pull/75
  • @matqr made their first contribution in https://github.com/koito19960406/ZenSVI/pull/82

Full Changelog: https://github.com/koito19960406/ZenSVI/compare/v0.12.0...v.1.0.0

- Python
Published by koito19960406 about 1 year ago

zensvi - New classification model: PlacePulse perception prediction

ClassifierPerception added: New module for PlacePulse perception score prediction. Usage: from zensvi.cv import ClassifierPerception. You can set perception_study like this: ClassifierPerception(perception_study="safer"). Other options are "livelier", "wealthier", "more beautiful", "more boring", and "more depressing"

Many thanks for your contribution! @seshing

- Python
Published by koito19960406 almost 2 years ago

zensvi - New sub-packages: Metadata, various Claissifers, and DepthEstimator

  • Metadata Module Added: New module for metadata processing. Usage: from zensvi import metadata.
  • ClassifierGlare, ClassifierLighting, ClassifierPanorama, ClassifierPlatform, ClassifierQuality, ClassifierReflection, ClassifierViewDirection, ClassifierWeather Introduced: New classifiers for various image quality attributes. Usage: from zensvi.cv import ClassifierGlare, ClassifierLighting, ClassifierPanorama, ClassifierPlatform, ClassifierQuality, ClassifierReflection, ClassifierViewDirection, ClassifierWeather.
  • DepthEstimator Introduced: New module for depth estimation. Usage: from zensvi.cv import DepthEstimator.

- Python
Published by koito19960406 almost 2 years ago

zensvi - New modules: visualization, cv.low_level, and cv.classification

In this new minor version, three main changes were introduced:

  1. Addition of visualization module: you can import like this from zensvi import vialualization to plot maps (visualization.plot_map) and images (visualization.plot_image)
  2. Addition of lowlovel in cv module: you can import like this `from zensvi.cv import getlowlevelfeaturesto quantify low level features (getlowlevel_features`)
  3. Addition of ClassifierPlaces365: you can import like this from zensvi.cv import ClassifierPlaces365. This is based on the original model (https://github.com/CSAILVision/places365).

- Python
Published by koito19960406 almost 2 years ago

zensvi - Improved downloading functions

The following changes have been made:

  1. MLYDownloader: The efficiency of downloading improved a lot by internally using polygon to download, rather than poiints. But the user needs to set buffer above 0 when the input is either lat and lon, input_csv_file, or Point/Line in input_shp_file.
  2. GSVDownloader: The data cleaning process has been implemented, and there will be no more overlaps and dark areas in older images.
  3. GSVDownloader: h_tiles and v_tiles arguments in download_svi function have been deprecated because now the function automatically calculates them internally.
  4. MLYDownloader & GSVDownloader: metadata_only argument has been added to download_svi function to allow users to only download metadata of SVIs (e.g., id, time, location, etc).

- Python
Published by koito19960406 about 2 years ago

zensvi - Flexible image transformation from panorama to perspective

I added theta as an argument to transform_images() function in ImageTransformer class, which will allow users to modify the angle of views without affecting the field of view. So for example, this new argument allows users to do the following (which wasn't allowed to do in previous versions): Getting 120 degree FOV perspective from each of 90 angle (theta) of the panorama image.

- Python
Published by koito19960406 over 2 years ago

zensvi - Flexible batch size when downloading SVI

Having a large number of image files in a single folder can cause some issues, so I introduced sub-folders when downloading SVI with an optional batch_size argument in download_svi function in GSVDownloader and MLYDownloader

- Python
Published by koito19960406 over 2 years ago

zensvi - Panoptic segmentation task and csv output format

This new release enables conducting panoptic segmentation (i.e. semantic + instance segmentation) in addition to the normal semantic segmentation.

The initialization of Segmenter also changed by deprecating model argument and introducing task argument instead. Also, to accommodate the output of panoptic segmentation of "labelcounts.csv" and/or "labelcounts.json" files, we changed dirpixelratiooutput to **dirsegmentationsummaryoutput**. Moreover, you can now also choose the wide/long format of csv output by changing csv_format argument to either long or wide. So you can use cityscapes dataset and conduct panoptic segmentation by running the following code:

```

initialize the segmenter

dataset="cityscapes" task="panoptic" segmenter = Segmenter(dataset=dataset, task=task)

set arguments

dirinput = "path/to/input" dirimageoutput = "path/to/imageoutput" dirsegmentationsummaryoutput = "path/to/pixelratiooutput" saveimageoptions = ["segmentedimage", "blendimage"] pixelratiosaveformat = ["csv", "json"] csvformat = "long" # "long" or "wide" segmenter.segment(dirinput, dirimageoutput = dirimageoutput, dirsegmentationsummaryoutput = dirpixelratiooutput, saveimageoptions = saveimageoptions, pixelratiosaveformat = pixelratiosaveformat, csvformat=csvformat)

```

- Python
Published by koito19960406 over 2 years ago

zensvi - New additional arguments for download_svi

added new arguments for download_svi function from all the Downloader classes:

cropped: if True, only the upper half of the image is saved full: if True, the full image is saved. If False, the image tiles are separately saved (* GSVDownloader only) OSMnx argument: network_type OSMnx argument: custom_filter

- Python
Published by koito19960406 over 2 years ago

zensvi - More flexible image transformation methods

This version has two changes in the ImageTransformer.transform_images function.

  1. FOV and aspects optional arguments. FOV controls the angle, which is used to split 360-degree panorama image. The default is 90 degreesFOV=90. aspects controls the height and width ration of the panorama image. The default is (9, 16) aspects=(9, 16).
  2. Fish eye transformation methods as options in style_list. There are now ["perspective", "equidistant_fisheye", "orthographic_fisheye", "stereographic_fisheye", "equisolid_fisheye"] to choose from. Please refer to the Wikipedia page here

- Python
Published by koito19960406 almost 3 years ago

zensvi - Introduced Downloaders for Google Street View and Mapillary

Created new Downloaders named GSVDownloader and MLYDownloder for Google Street View and Mapillary, respectively. The older StreetViewDownloader is now deprecated. Below are some example codes. downloader = GSVDownloader(gsv_api_key = gsv_api_key, distance=1, grid = False, grid_size = 1) downloader.download_svi("path/to/output", lat=52.078663, lon=4.313877, augment_metadata=True) downloader = MLYDownloader(mly_api_key=YOUR_MAPILLARY_API_TOKEN) downloader.download_svi(dir_output = "tests/data/output", lat=1.276095, lon=103.792547, radius=50, # you can input keyword arguments for mapillary Python SDK functions too organization_id=int(org_id) )

- Python
Published by koito19960406 almost 3 years ago

zensvi - New input called input_place_name

This new minor version comes with a new argument in downloadgsv function called `inputplace_name`, which takes a string value to download GSV within the boundary of the place. You can use it as follows:

from zensvi.download import StreetViewDownloader downloader = StreetViewDownloader() downloader.download_gsv("path/to/output_directory", input_place_name="Singapore")

- Python
Published by koito19960406 almost 3 years ago

zensvi - Improved spatial processing speed and ID columns as a new input

This version sped up the spatial post-processing and enabled users to input ID columns of csv and shp files to keep track of IDs of the input file and panorama IDs.

- Python
Published by koito19960406 almost 3 years ago