Saliency-Driven Versatile Video Coding for Neural Object Detection

TL;DR

This paper introduces a saliency-driven video coding framework using VVC, YOLO, and Mask R-CNN, achieving up to 29% bitrate savings while maintaining detection accuracy for machine vision tasks.

Contribution

It presents a novel saliency-driven coding approach for machine vision that combines YOLO-based salient region detection with VVC, optimizing bitrate efficiency.

Findings

Up to 29% bitrate savings at the same detection accuracy.

Effective saliency detection using YOLO outperforms traditional methods.

Framework compatible with real-time object detection and segmentation.

Abstract

Saliency-driven image and video coding for humans has gained importance in the recent past. In this paper, we propose such a saliency-driven coding framework for the video coding for machines task using the latest video coding standard Versatile Video Coding (VVC). To determine the salient regions before encoding, we employ the real-time-capable object detection network You Only Look Once~(YOLO) in combination with a novel decision criterion. To measure the coding quality for a machine, the state-of-the-art object segmentation network Mask R-CNN was applied to the decoded frame. From extensive simulations we find that, compared to the reference VVC with a constant quality, up to 29 % of bitrate can be saved with the same detection accuracy at the decoder side by applying the proposed saliency-driven framework. Besides, we compare YOLO against other, more traditional saliency detection methods.