Computer vision system and method

This invention describes a method for breaking down an image into meaningful segments, such as identifying different objects or regions. It works by taking the original image and a smaller, lower-resolution copy of it. Both images are processed independently by their own neural networks, which use an efficient technique called separable convolution. The network processing the original image has fewer layers than the one processing the lower-resolution image. The results from these two networks are then combined by simply adding them together at a single stage to produce the final segmented image.

Why it matters: Filed when deep learning models were often computationally heavy. Advances in efficient neural network architectures and hardware make multi-resolution segmentation with separable convolutions more practical for real-time applications.

Status: Not in forceListed as no longer active. The specific reason is not in the record we hold.
How hard to build: SpecializedMachine learning software, neural network architecture

3 AI reinvention directions for this patent

See all three

AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.

Reinvent this with AI

Where this one came from

Every morning we go through millions of abandoned patents and pick out 15 worth your time. This was one of today’s.

This one

+10