Acotate

It would be very nice if there were already AI models for every possible object class that find objects automatically. However, this is not the case for many object classes. For these, a large amount of data is needed to detect them as well. Acotate tries to make this process of time-consuming manual annotation as pleasant as possible - especially with multisensory data, the effort per camera perspective would be enormous.

Acotate offers the possibility to interpolate between frames. To do this, at least two frames must be annotated according to the consistency specifications, with any number of empty frames between them. The smaller the distance between the two frames is selected, the more accurate the result will be. The empty frames are then filled with the same objects that exist in both manually annotated frames.

Not yet. We at Accotate want to help make the annotation of new object classes as easy as possible. There are many tools out there that can apply pre-trained models to data. We are aimed at those who need to go one step further - for those objects that cannot yet be found automatically.

Yes - Acotate provides on-demand labeling services such as creating a labeling guide to streamline the labeling process. However, Acotate does not employ labelers. The Acotate platform is designed to reduce the time it takes to label each asset, so our labelers are focused on completing tasks faster and on increasing the value of each label. Rather than paying for labeling work on a fixed label-cost model, you only pay for the time when labeling occurs. Labeling hours can either be pre-paid in blocks of 40 hours or billed monthly for annually contracted customers.