18482463. DEEP GEOMETRIC MODEL FITTING simplified abstract (Intel Corporation)

From WikiPatents
Jump to navigation Jump to search

DEEP GEOMETRIC MODEL FITTING

Organization Name

Intel Corporation

Inventor(s)

Rene Ranftl of Munich (DE)

Vladlen Koltun of Santa Clara CA (US)

DEEP GEOMETRIC MODEL FITTING - A simplified explanation of the abstract

This abstract first appeared for US patent application 18482463 titled 'DEEP GEOMETRIC MODEL FITTING

Simplified Explanation

The technology described in the patent application involves the use of two neural networks to generate and update model weights based on input data and residual data, ultimately outputting a geometric model of the input data.

  • The first neural network generates an initial set of model weights based on input data.
  • The second neural network iteratively generates an updated set of model weights based on residual data associated with the initial set of model weights and the input data.
  • The technology outputs a geometric model of the input data based on the updated set of model weights.
  • The first and second neural networks reduce the dependence of the geometric model on the number of data points in the input data.

Potential Applications

This technology could be applied in various fields such as image recognition, pattern recognition, and data analysis.

Problems Solved

This technology helps in improving the accuracy and efficiency of generating geometric models from input data.

Benefits

The benefits of this technology include improved model accuracy, reduced dependency on the number of data points, and enhanced performance in generating geometric models.

Potential Commercial Applications

One potential commercial application of this technology could be in the development of advanced data analysis tools for industries such as healthcare, finance, and manufacturing.

Possible Prior Art

Prior art in the field of neural networks and geometric modeling may include similar techniques used in image processing and machine learning algorithms.

Unanswered Questions

How does this technology compare to existing methods for generating geometric models from input data?

This technology uses a unique approach involving two neural networks to iteratively update model weights, potentially leading to more accurate geometric models. However, it would be beneficial to compare its performance and efficiency with traditional methods.

What are the limitations or challenges of implementing this technology in real-world applications?

While the technology shows promise in generating geometric models, there may be challenges in scaling it for large datasets or integrating it into existing systems. Understanding these limitations can help in optimizing its practical use.


Original Abstract Submitted

Systems, apparatuses and methods may provide for technology that generates, by a first neural network, an initial set of model weights based on input data and iteratively generates, by a second neural network, an updated set of model weights based on residual data associated with the initial set of model weights and the input data. Additionally, the technology may output a geometric model of the input data based on the updated set of model weights. In one example, the first neural network and the second neural network reduce the dependence of the geometric model on the number of data points in the input data.