Multi-resolution Rescored ByteTrack for Video Object Detection on Ultra-low-power Embedded Systems
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866929317410242560 |
|---|---|
| author | Bompani, Luca Rusci, Manuele Palossi, Daniele Conti, Francesco Benini, Luca |
| author_facet | Bompani, Luca Rusci, Manuele Palossi, Daniele Conti, Francesco Benini, Luca |
| contents | This paper introduces Multi-Resolution Rescored Byte-Track (MR2-ByteTrack), a novel video object detection framework for ultra-low-power embedded processors. This method reduces the average compute load of an off-the-shelf Deep Neural Network (DNN) based object detector by up to 2.25$\times$ by alternating the processing of high-resolution images (320$\times$320 pixels) with multiple down-sized frames (192$\times$192 pixels). To tackle the accuracy degradation due to the reduced image input size, MR2-ByteTrack correlates the output detections over time using the ByteTrack tracker and corrects potential misclassification using a novel probabilistic Rescore algorithm. By interleaving two down-sized images for every high-resolution one as the input of different state-of-the-art DNN object detectors with our MR2-ByteTrack, we demonstrate an average accuracy increase of 2.16% and a latency reduction of 43% on the GAP9 microcontroller compared to a baseline frame-by-frame inference scheme using exclusively full-resolution images. Code available at: https://github.com/Bomps4/Multi_Resolution_Rescored_ByteTrack |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2404_11488 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Multi-resolution Rescored ByteTrack for Video Object Detection on Ultra-low-power Embedded Systems Bompani, Luca Rusci, Manuele Palossi, Daniele Conti, Francesco Benini, Luca Computer Vision and Pattern Recognition Artificial Intelligence I.4 This paper introduces Multi-Resolution Rescored Byte-Track (MR2-ByteTrack), a novel video object detection framework for ultra-low-power embedded processors. This method reduces the average compute load of an off-the-shelf Deep Neural Network (DNN) based object detector by up to 2.25$\times$ by alternating the processing of high-resolution images (320$\times$320 pixels) with multiple down-sized frames (192$\times$192 pixels). To tackle the accuracy degradation due to the reduced image input size, MR2-ByteTrack correlates the output detections over time using the ByteTrack tracker and corrects potential misclassification using a novel probabilistic Rescore algorithm. By interleaving two down-sized images for every high-resolution one as the input of different state-of-the-art DNN object detectors with our MR2-ByteTrack, we demonstrate an average accuracy increase of 2.16% and a latency reduction of 43% on the GAP9 microcontroller compared to a baseline frame-by-frame inference scheme using exclusively full-resolution images. Code available at: https://github.com/Bomps4/Multi_Resolution_Rescored_ByteTrack |
| title | Multi-resolution Rescored ByteTrack for Video Object Detection on Ultra-low-power Embedded Systems |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence I.4 |
| url | https://arxiv.org/abs/2404.11488 |