Released Versions of Capto

Added RGB image publishing from cameras.
Introduced more tools for calling APIs from the frontend.
Made the config files configurable directly from the frontend.
Fixed UR controls to run smoother.
Added an API function to check if the compartment sizes are equal using models for checking compartment masks.
Integrated Kinova robots and general URDF file importation into the system.

New Hardware Support: Added support for Yaskawa GP25 Robot, Kinova Robot, and KLT Bin Gripper.
Bin Handling: Improved bin detection, post-processing, and added bin rotation handling.
Improved depth filtering.
Introduced a new algorithm for picking and placing totes.
Added eye-in-hand calibration.
Enhanced two-finger grasping, improved collision checking, and added new configurations.
Added an event publisher for API calls.
Fixed issues with camera, ROS, and dockerization.

Added API calls to get_next_object_size and get_object_sizes for retrieving object sizes.
Upgraded "fixed_item_to_bin" place type to allow sampling place poses on a specific axis via the configuration sample_on_axis.
Added APIs for calling VQA prompts and comparing total object volume with compartment volume.
Improved the placing function.
Modified the publisher to properly handle errors in the PredictionUploader, ensuring correct operation.

Implemented a new structure for barcode scanning.
Updated calibrated data immediately after calibration.
Introduced popovers to improve frontend responsiveness.
Added bin detection to the compartment API call.
Uploaded additional necessary data, including time per execution, to improve functionality.

Added new placing types: fixed_item_to_bin and tight_packaging.
Fixed visualization issues.
Resolved all configuration issues, allowing direct configuration in the frontend.
Automatically updated calibration pose for the camera in the frontend.
Integrated RealSense D405 for enhanced sensor capabilities.
Integrated various bin detection functions: bin detection, bin sanity check, bin move check, etc.
Improved data upload to the server using blobs.
Enabled automatic retrieval of calibration files for UR robots.
Introduced functionality to reduce bag size, ensuring compatibility with the API.

Introduced soft placing for objects that do not fit the standard criteria.
Added the ability to configure the operational area.
Increased the number of placing angles for specific scenarios.
Incorporated fixed object sizes for more precise handling.
Implemented new grasp generators with a sampling, coarse check, and fine check workflow.
Significantly improved the performance of the two-finger gripper.
Added API calls to check the number of objects in the bin.
Introduced an API call to verify if all objects in the pick bin are suitable for placement.