17973370. SALIENCY-BASED INPUT RESAMPLING FOR EFFICIENT OBJECT DETECTION simplified abstract (QUALCOMM Incorporated)

From WikiPatents
Jump to navigation Jump to search

SALIENCY-BASED INPUT RESAMPLING FOR EFFICIENT OBJECT DETECTION

Organization Name

QUALCOMM Incorporated

Inventor(s)

Babak Ehteshami Bejnordi of Amsterdam (NL)

Amir Ghodrati of Amsterdam (NL)

Fatih Murat Porikli of San Diego CA (US)

Amirhossein Habibian of Amsterdam (NL)

SALIENCY-BASED INPUT RESAMPLING FOR EFFICIENT OBJECT DETECTION - A simplified explanation of the abstract

This abstract first appeared for US patent application 17973370 titled 'SALIENCY-BASED INPUT RESAMPLING FOR EFFICIENT OBJECT DETECTION

Simplified Explanation

The abstract describes a method of video processing using an artificial neural network (ANN). Here are the key points:

  • The method involves receiving a video with two frames.
  • A saliency map is generated based on the first frame, which highlights the most important areas of the frame.
  • The second frame is then sampled based on the saliency map.
  • The second frame is divided into two portions, with each portion sampled at a different resolution.
  • A resampled second frame is generated based on the sampling.
  • The resampled second frame is processed to determine an inference associated with the video.

Potential Applications

  • Video analysis and processing in various fields such as surveillance, entertainment, and sports.
  • Enhancing video compression techniques by focusing on important areas of the video.
  • Improving video editing and post-production processes by automatically identifying significant parts of the footage.

Problems Solved

  • Efficiently processing videos by focusing on the most relevant parts, reducing computational resources.
  • Enhancing video analysis by considering the saliency of frames and prioritizing important information.
  • Improving the accuracy of inferences and predictions made from video data.

Benefits

  • More efficient video processing, reducing computational requirements and time.
  • Improved accuracy in analyzing and understanding video content.
  • Enhanced video compression and editing capabilities, leading to better quality and more engaging content.


Original Abstract Submitted

A processor-implemented method of video processing using includes receiving, via an artificial neural network (ANN), a video including a first frame and a second frame. A saliency map is generated based on the first frame of the video. The second frame of the video is sampled based on the saliency map. A first portion of the second frame is sampled at a first resolution and a second portion of the second frame is sampled at a second resolution. The first resolution is different than the second resolution. A resampled second frame is generated based on the sampling of the second frame. The resampled second frame is processed to determine an inference associated with the video.

Retrieved from "http://wikipatents.org/index.php?title=17973370._SALIENCY-BASED_INPUT_RESAMPLING_FOR_EFFICIENT_OBJECT_DETECTION_simplified_abstract_(QUALCOMM_Incorporated)&oldid=23299"