apple/ml-hypersim

We implement a simple tonemapping operator in hypersim/code/python/tools/scene_generate_images_tonemap.py.Lossy preview images that are useful for debugging are stored at the following locations within each ZIP file:Each camera trajectory is stored as a dense list of camera poses at the following location within each ZIP file:We recommend browsing through hypersim/code/python/tools/scene_generate_images_bounding_box.py to understand our camera pose conventions. The Hypersim Low-Level Toolkit consists of the following Python command-line tools.The Hypersim High-Level Toolkit consists of the following Python command-line tools.The Hypersim High-Level Toolkit also includes the Hypersim Scene Annotation Tool executable, which is located in the hypersim/code/cpp/bin directory, and can be launched from the command-line as follows.The following tutorial examples demonstrate the functionality in the Hypersim Toolkit.00_empty_scene In this tutorial example, we use the Hypersim Low-Level Toolkit to add a camera trajectory and a collection of textured quads to a V-Ray scene.01_marketplace_dataset In this tutorial example, we use the Hypersim High-Level Toolkit to export and manipulate a scene downloaded from a content marketplace. We recommend making a note of the absolute Windows path to the dataset, because you will need to supply it whenever a subsequent pipeline step requires the dataset_dir_when_rendering argument.If you are generating data on portable hard drives, we recommend running our pipeline in batches of 10 volumes at a time (i.e., roughly 100 scenes at a time), and storing each batch on its own 4TB drive. We include the argument --scene_names ai_00* in our instructions below.When preparing the Hypersim Dataset, we chose to manually exclude some scenes and automatically generated camera trajectories. There is no harm in running our pipeline for these scenes, but it is possible to save a bit of time and money by not rendering images for these manually excluded scenes and camera trajectories.The camera trajectories we manually excluded from our dataset are listed in hypersim/evermotion_dataset/analysis/metadata_camera_trajectories.csv. So, you can use our automatic pipeline to generate instance-level semantic segmentation images without needing to manually annotate any scenes.We include the cost of rendering each image in our dataset in hypersim/evermotion_dataset/analysis/metadata_rendering_tasks.csv. To compute the total cost of the image frame.0000 in the camera trajectory cam_00 in the scene ai_001_001, we add up the vray_cost_dollars and cloud_cost_dollars columns for the rows where job_name == {ai_001_001@scene_cam_00_geometry, ai_001_001@scene_cam_00_pre, ai_001_001@scene_cam_00_final} and task_id == 0.To process the first batch of scenes (Volumes 1-9) in the Hypersim Dataset, we execute the following pipeline steps.