AppendInstanceDiffusionTracking

The AppendInstanceDiffusionTracking node receives two inputs of type TRACKING. Its internal implementation merges the tracking data from the second input into the data of the first input.

The AppendInstanceDiffusionTracking node is an auxiliary tool for processing InstanceDiffusion tracking data. Its core function is to merge two sets of tracking data into one and concatenate the corresponding prompt texts, preparing the data for subsequent instance segmentation and generation steps.

Node Functionality

This node receives two inputs of type TRACKING. Its internal implementation merges the tracking data from the second input into the data of the first input. If categories with the same name exist in both inputs, the node will merge the specific instance data (class IDs) under that category without overwriting or throwing an error. Additionally, the two optional text prompts will be merged into a single comma-separated string for output.

Node Parameter Description - AppendInstanceDiffusionTracking

Connection Inputs (Inputs)

Parameter Name	Data Type	Required	Default Value	Value Range/Options	Description
`tracking_1`	TRACKING	Yes	-	-	The first set of instance tracking data, which serves as the base for the merged data.
`tracking_2`	TRACKING	Yes	-	-	The second set of instance tracking data, whose content will be merged into the first set.

Control Parameters (Parameters)

Parameter Name	Data Type	Required	Default Value	Value Range/Options	Description
`prompt_1`	STRING	No	""	-	Optional prompt text to be concatenated.
`prompt_2`	STRING	No	""	-	Optional prompt text to be concatenated.

Output

Parameter Name	Data Type	Description
tracking	TRACKING	The merged instance tracking data.
prompt	STRING	A string formed by concatenating the two input prompts.

Usage Scenarios

In practical applications, you can use this node when your workflow needs to integrate instance detection results from different sources (e.g., different video frames or image regions). For example, you can first perform object detection and tracking on the first half and the second half of a video separately, then merge the tracking results from both parts using this node to obtain a unified dataset containing all instance information for the entire video. This dataset can then be fed into the InstanceDiffusion node for image generation.

Notes

Please note that this node is primarily designed to handle the specific TRACKING data structure required by the InstanceDiffusion extension. The input data format must comply with the requirements; otherwise, it may cause errors in downstream nodes.