BROWSE
Detail
Conference Paper
Pixel-Unshuffled Multi-level Feature Map Compression for FCVCM
Cited 1 time in 
Share 
Share
- Authors
- Younhee Kim, Se-Yoon Jeong, Jooyoung Lee, Jongseok Lee, Minsub Kim
- Issue Date
- 2023-12
- Citation
- International Conference on Visual Communications and Image Processing (VCIP) 2023, pp.1-4
- Language
- English
- Type
- Conference Paper
- Abstract
- The feature compression process for machine tasks involves several steps: feeding a video into the task network, extracting intermediate feature maps, compressing these maps into a bitstream on the client side, transmitting the bitstream to a resource-rich server, decoding it, and ultimately completing the specific machine task. In this paper, we present a multi-level feature compression method designed for machine tasks. We introduce an efficient and effective feature reshaping and merging module within the PCA-based feature coding scheme. This module utilizes pixel-unshuffled operations to reshape the multi-level features, merges them into a single map, and then performs a transformation. Our proposed method achieves a BD-rate gain of 49.69% and 66.3% in comparison to the previous computational low cost PCA-based feature coding method for object detection and instance segmentation tasks, respectively.
- KSP Keywords
- Client side, Coding method, Compression method, Compression process, Feature compression, Feature map, Low-cost, Multi-level feature, PCA-based, Task Network, coding scheme