Method and Apparatus for Analysing an Image
AssigneeSiemens Aktiengesellschaft
InventorsSanjukta Ghosh, Peter Amon, Andreas Hutter
Life of this patent
Estimating the term…
Abstract
The present embodiments relate to analysing an image. An artificial deep neural net is pre-trained to classify images into a hierarchical system of multiple hierarchical classes. The pre-trained neural net is then adapted for one specific class, wherein the specific class is lower in the hierarchical system than an actual class of the image. The image is then processed by a forward pass through the adapted neural net to generate a processing result. An image processing algorithm is then used to analyse the processing result focused on features corresponding to the specific class.
Claims19
A method for analysing an image, the method comprising:
- providing an artificial deep neural net pre-trained to classify images into a hierarchical system of multiple hierarchical classes,
- adapting the pre-trained neural net for one specific class of the multiple hierarchical classes, wherein the specific class is lower in the hierarchical system than an actual class of the image,
- processing the image by a forward pass through the adapted neural net, the processing generating a processing result, and
- analysing the processing result focused on features corresponding to the specific class using an image processing algorithm.
An apparatus for analysing an image, the apparatus comprising:
- an artificial deep neural net pre-trained to classify images into a hierarchical system of multiple hierarchical classes, and
- a separate image processor, wherein the processor is configured to:
- process the image by a forward pass through an adapted neural net to generate a processing result, wherein the adapted neural net is adapted from the pre-trained neural net for one specific class of the multiple hierarchical classes, wherein the specific class is lower in the hierarchical system than an actual class of the image, and
- analyse the processing result focused on features corresponding to the specific class using an image processing algorithm.
Dependencies are inferred from the claim text.