Nilvana Vision Studio - Manual Annotation
Please follow the instructions to get started manual annotation with nilvana vision studio.
Multi-Person Real-Time Manual Annotation
Annotation is often the most labor-intensive and time-consuming task in the model development process. Vision studio not only provides user-friendly machine annotation (if you are unsure with how to use machine annotation, please refer to "Nilvana Vision Studio - Model Training") to assist you in completing most of the annotation operations, but also provides a multi-person real-time manual annotation tool to help you master the status of all annotations in real-time and save time in transferring and merging annotation data. In addition, the advantage of multi-person real-time annotation lies in the ability to discuss remotely and immediately correct the results of each other's annotations to improve the quality of annotation.
We continuously test and optimize the annotation process to create an annotation tool with the best user experience design. If a member is working on an annotation, the manual annotation page will display the member's current status and the corresponding box line.
Individual Image Annotation Version Records
For each image, the editing history of each member in the image is saved, so that you can easily check or restore the image to the previous annotated version.
Adjustment of Imported Annotated Data
In addition to supporting the import of annotated data into the dataset, you can also adjust the imported annotated data through the manual annotation tool, and continue to add and delete images to this dataset, so that the quality of the dataset can be continuously improved.
Hotkeys to Speed Up Annotation
Utilize hotkeys to facilitate smoother manual annotation. For example, when you add a new annotation box, you can use the numeric keys to quickly switch the type of the annotation box; quickly switch between various annotation tools; and move the annotation box by small margins, so that you can annotate objects more accurately. More useful functions await your discovery, come enjoy the convenient and efficient annotation experience!
Note: Dataset comes from the mask facial database developed by the Eden Social Welfare Foundation's "Digital Data Processing Sheltered Workshop".