Object Detection Post Processing Accelerator Based on Co-Design of Hardware and Software

Guardado en:
Detalles Bibliográficos
Publicado en:Information vol. 16, no. 1 (2025), p. 63
Autor principal: Yang, Dengtian
Otros Autores: Chen, Lan, Hao, Xiaoran, Zhang, Yiheng
Publicado:
MDPI AG
Materias:
Acceso en línea:Citation/Abstract
Full Text + Graphics
Full Text - PDF
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!

MARC

LEADER 00000nab a2200000uu 4500
001 3159490312
003 UK-CbPIL
022 |a 2078-2489 
024 7 |a 10.3390/info16010063  |2 doi 
035 |a 3159490312 
045 2 |b d20250101  |b d20251231 
084 |a 231474  |2 nlm 
100 1 |a Yang, Dengtian  |u Institute of Microelectronics of the Chinese Academy of Sciences, Beijing 100029, China; <email>yangdengtian@ime.ac.cn</email> (D.Y.); <email>haoxiaoran@ime.ac.cn</email> (X.H.); <email>zhangyiheng@ime.ac.cn</email> (Y.Z.); University of Chinese Academy of Sciences, Beijing 100049, China 
245 1 |a Object Detection Post Processing Accelerator Based on Co-Design of Hardware and Software 
260 |b MDPI AG  |c 2025 
513 |a Journal Article 
520 3 |a Deep learning significantly advances object detection. Post processes, a critical component of this process, select valid bounding boxes to represent the true targets during inference and assign boxes and labels to these objects during training to optimize the loss function. However, post processes constitute a substantial portion of the total processing time for a single image. This inefficiency primarily arises from the extensive Intersection over Union (IoU) calculations required between numerous redundant bounding boxes in post processing algorithms. To reduce these redundant IoU calculations, we introduce a classification prioritization strategy during both training and inference post processes. Additionally, post processes involve sorting operations that contribute to their inefficiency. To minimize unnecessary comparisons in Top-K sorting, we have improved the bitonic sorter by developing a hybrid bitonic algorithm. These improvements have effectively accelerated the post processing. Given the similarities between the training and inference post processes, we unify four typical post processing algorithms and design a hardware accelerator based on this framework. Our accelerator achieves at least 7.55 times the speed in inference post processing compared to that of recent accelerators. When compared to the RTX 2080 Ti system, our proposed accelerator offers at least 21.93 times the speed for the training post process and 19.89 times for the inference post process, thereby significantly enhancing the efficiency of loss function minimization. 
653 |a Software 
653 |a Accuracy 
653 |a Co-design 
653 |a Deep learning 
653 |a Back propagation 
653 |a Hardware 
653 |a Sensors 
653 |a Boxes 
653 |a Classification 
653 |a Inference 
653 |a Target detection 
653 |a Algorithms 
653 |a Object recognition 
653 |a Machine learning 
653 |a Critical components 
653 |a Design optimization 
653 |a Sorting algorithms 
653 |a Efficiency 
700 1 |a Chen, Lan  |u Institute of Microelectronics of the Chinese Academy of Sciences, Beijing 100029, China; <email>yangdengtian@ime.ac.cn</email> (D.Y.); <email>haoxiaoran@ime.ac.cn</email> (X.H.); <email>zhangyiheng@ime.ac.cn</email> (Y.Z.) 
700 1 |a Hao, Xiaoran  |u Institute of Microelectronics of the Chinese Academy of Sciences, Beijing 100029, China; <email>yangdengtian@ime.ac.cn</email> (D.Y.); <email>haoxiaoran@ime.ac.cn</email> (X.H.); <email>zhangyiheng@ime.ac.cn</email> (Y.Z.) 
700 1 |a Zhang, Yiheng  |u Institute of Microelectronics of the Chinese Academy of Sciences, Beijing 100029, China; <email>yangdengtian@ime.ac.cn</email> (D.Y.); <email>haoxiaoran@ime.ac.cn</email> (X.H.); <email>zhangyiheng@ime.ac.cn</email> (Y.Z.) 
773 0 |t Information  |g vol. 16, no. 1 (2025), p. 63 
786 0 |d ProQuest  |t Advanced Technologies & Aerospace Database 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/3159490312/abstract/embedded/L8HZQI7Z43R0LA5T?source=fedsrch 
856 4 0 |3 Full Text + Graphics  |u https://www.proquest.com/docview/3159490312/fulltextwithgraphics/embedded/L8HZQI7Z43R0LA5T?source=fedsrch 
856 4 0 |3 Full Text - PDF  |u https://www.proquest.com/docview/3159490312/fulltextPDF/embedded/L8HZQI7Z43R0LA5T?source=fedsrch