Title :
A classification centric quantizer for efficient encoding of predictive feature errors
Author :
Chen, Scott Deeann ; Moulin, Pierre
Author_Institution :
Dept. of Electr. & Comput. Eng., Univ. of Illinois at Urbana-Champaign, Urbana, IL, USA
Abstract :
We design a joint compression and classification system that optimizes visual fidelity and classification accuracy under a bit rate constraint. We propose a classification centric quantizer (CCQ) whose parameters are learned from labeled training data. We apply and evaluate the CCQ on a scene classification problem and compare the results to previous work.
Keywords :
image classification; image coding; learning (artificial intelligence); CCQ; bit rate constraint; classification accuracy; classification centric quantizer; labeled training data; predictive feature error encoding; scene classification problem; visual fidelity optimization; Accuracy; Feature extraction; Image coding; Kernel; PSNR; Quantization (signal); Support vector machines;
Conference_Titel :
Signals, Systems and Computers, 2014 48th Asilomar Conference on
Print_ISBN :
978-1-4799-8295-0
DOI :
10.1109/ACSSC.2014.7094844
