17434729. COMPUTER-IMPLEMENTED IMAGE-PROCESSING METHOD, IMAGE-ENHANCING CONVOLUTIONAL NEURAL NETWORK, AND COMPUTER PRODUCT simplified abstract (BOE Technology Group Co., Ltd.)

From WikiPatents
Jump to navigation Jump to search

COMPUTER-IMPLEMENTED IMAGE-PROCESSING METHOD, IMAGE-ENHANCING CONVOLUTIONAL NEURAL NETWORK, AND COMPUTER PRODUCT

Organization Name

BOE Technology Group Co., Ltd.

Inventor(s)

Dan Zhu of Beijing (CN)

Guannan Chen of Beijing (CN)

Ran Duan of Beijing (CN)

COMPUTER-IMPLEMENTED IMAGE-PROCESSING METHOD, IMAGE-ENHANCING CONVOLUTIONAL NEURAL NETWORK, AND COMPUTER PRODUCT - A simplified explanation of the abstract

This abstract first appeared for US patent application 17434729 titled 'COMPUTER-IMPLEMENTED IMAGE-PROCESSING METHOD, IMAGE-ENHANCING CONVOLUTIONAL NEURAL NETWORK, AND COMPUTER PRODUCT

Simplified Explanation

The abstract describes a computer-implemented image-processing method that involves enhancing the sharpness of an image using a convolutional neural network. The method includes obtaining a pair of training samples, one with a lower degree of sharpness (training image) and the other with a higher degree of sharpness (reference image). The training image is inputted into the image-enhancing convolutional neural network to generate an enhanced image. The enhanced image is then processed by an edge detector to generate multiple edge maps. The same process is repeated with the reference image. Finally, parameters in the neural network are adjusted to minimize the losses in both the generated edge maps and the reference image.

  • Obtaining a pair of training samples with different degrees of sharpness
  • Inputting the training image into a convolutional neural network for enhancement
  • Using an edge detector to generate edge maps from the enhanced image and reference image
  • Tuning parameters in the neural network to minimize losses in the edge maps and reference image

Potential Applications

  • Image enhancement for various purposes such as photography, medical imaging, and surveillance
  • Improving the quality and clarity of images in real-time video processing
  • Enhancing the sharpness of images in computer vision applications like object detection and recognition

Problems Solved

  • Overcoming the limitations of traditional image enhancement techniques by using a convolutional neural network
  • Addressing the challenge of enhancing image sharpness while preserving image content and reducing artifacts
  • Providing a method to automatically adjust parameters in the neural network for optimal image enhancement

Benefits

  • Improved image sharpness and clarity for better visual perception and analysis
  • Automation of the image enhancement process, reducing the need for manual adjustments
  • Potential for real-time image enhancement in various applications


Original Abstract Submitted

A computer-implemented image-processing method is provided. The computer-implemented image-processing method includes obtaining a pair of training samples including a training image having a first degree of sharpness and a reference image having a second degree of sharpness, the second degree greater than the first degree, at least portions of the training image and the reference image in a same pair having same contents; inputting the training image to the image-enhancing convolutional neural network to generate a training enhanced image; inputting the training enhanced image into an edge detector; generating, by the edge detector, a plurality of first edge maps; inputting the reference image into the edge detector; generating, by the edge detector, a plurality of second edge maps; and tuning parameters in the image-enhancing convolutional neural network to minimize at least the one or more first losses and a second loss.

Retrieved from "http://wikipatents.org/index.php?title=17434729._COMPUTER-IMPLEMENTED_IMAGE-PROCESSING_METHOD,_IMAGE-ENHANCING_CONVOLUTIONAL_NEURAL_NETWORK,_AND_COMPUTER_PRODUCT_simplified_abstract_(BOE_Technology_Group_Co.,_Ltd.)&oldid=18300"